From 6bf881a8ca453aa337be79c6945f1b97daf44039 Mon Sep 17 00:00:00 2001 From: ali Date: Mon, 20 Apr 2015 12:17:20 +0200 Subject: [PATCH 01/16] Update by Ali --- ACRONYMS.tex | 37 +++++++++ CHAPITRE_01.tex | 4 +- Thesis.tex | 8 +- Thesis.toc | 202 ++++++++++++++++++++++++------------------------ entete.tex | 7 +- 5 files changed, 151 insertions(+), 107 deletions(-) create mode 100644 ACRONYMS.tex diff --git a/ACRONYMS.tex b/ACRONYMS.tex new file mode 100644 index 0000000..c873dfd --- /dev/null +++ b/ACRONYMS.tex @@ -0,0 +1,37 @@ +\chapter*{abbreviations \markboth{abbreviations}{abbreviations}} +\label{chap} +\addcontentsline{toc}{chapter}{abbreviations} + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%% %% +%% abbreviations %% +%% %% +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + + +\begin{abbreviations} +\item[WSN] Wireless Sensor Network +\item[DILCO] Distributed Lifetime Coverage Optimization +\item[MuDiLCO] Multiround Distributed Lifetime Coverage Optimization +\item[PeCO] Perimeter-based Coverage Optimization +\item[DESK] Distributed Energy-efficient Scheduling for K-coverage +\item[GAF] Geographical Adaptive Fidelity +\item[PDA] Personal Digital Assistant +\item[WLAN] Wireless Local-Area Network +\item[MEMS] Micro-Electro-Mechanical Systems +\item[ADC] Analog to Digital Converters +\item[VCO] Voltage-Controlled Oscillator +\item[PLL] Phase-Locked Loop +\item[GPS] Global Positioning System +\item[OS] Operating System +\item[CMOS] Complementary Metal-Oxide-Silicon +\item[MAV] Micro Aerial Vehicle +\item[ECG] Electrocardiogram +\item[SCADA] Supervisory Control and Data Acquisition +\item[] +\item[] +\item[] +\item[] +\end{abbreviations} + + \ No newline at end of file diff --git a/CHAPITRE_01.tex b/CHAPITRE_01.tex index b6b4a4b..ec27e65 100644 --- a/CHAPITRE_01.tex +++ b/CHAPITRE_01.tex @@ -13,7 +13,7 @@ \label{ch1:sec:01} %The wireless networking has received more attention and fast growth in the last decade. In the last decade, wireless networking has became a major component of the global network infrastructure. -More precisely, the growing demand for the use of wireless applications and the continuous arrival of wireless devices such as portable computers, cellular phones, and personal digital assistants (PDAs) have led to develop different infrastructures of wireless networks. The wireless networks can be classified into two classes based on the network architecture~\cite{ref154,ref155}: Infrastructure-based networks that consist of a fixed network structure such as cellular networks and wireless local-area networks +More precisely, the growing demand for the use of wireless applications and the continuous arrival of wireless devices such as portable computers, cellular phones, and Personal Digital Assistants (PDAs) have led to develop different infrastructures of wireless networks. The wireless networks can be classified into two classes based on the network architecture~\cite{ref154,ref155}: Infrastructure-based networks that consist of a fixed network structure such as cellular networks and Wireless Local-Area Networks (WLANs); and Infrastructureless networks that are constructed dynamically by the cooperation of the wireless nodes in the network, where each node is capable of sending packets and taking decisions based on the network status. Examples of such type of networks include mobile ad hoc networks and wireless sensor networks. Figure~\ref{WNT} shows the taxonomy of wireless networks. \begin{figure}[h!] @@ -51,7 +51,7 @@ Figure~\ref{twsn} shows the components of a typical wireless sensor node~\cite{r \item \textbf{Computation Unit:} The main purpose of this unit is to manage and manipulate the instructions that are related to sensing, communication, and self-organization. This allows the sensor node to cooperate with other sensor nodes in order to perform the allocated sensing tasks. It is composed of a processor chip, an active short-term memory for storing the sensed data, an internal flash memory for storing program instructions, and an internal timer. -\item \textbf{Communication Unit:} It is responsible for all data transmission and reception done by the sensor node, which are performed by the transceiver circuitry. A transceiver circuit is composed of a mixer, frequency synthesizer, voltage-controlled oscillator (VCO), phase-locked loop (PLL), demodulator, and power amplifiers. All these components consume valuable power~\cite{ref19}. +\item \textbf{Communication Unit:} It is responsible for all data transmission and reception done by the sensor node, which are performed by the transceiver circuitry. A transceiver circuit is composed of a mixer, frequency synthesizer, Voltage-Controlled Oscillator (VCO), Phase-Locked Loop (PLL), demodulator, and power amplifiers. All these components consume valuable power~\cite{ref19}. \item \textbf{Power Unit:} This unit represents the most significant part of a sensor node. It supplies the other units by the needed power. diff --git a/Thesis.tex b/Thesis.tex index 2105500..e1bfe33 100644 --- a/Thesis.tex +++ b/Thesis.tex @@ -30,14 +30,18 @@ \listofalgorithms \addcontentsline{toc}{chapter}{List of Algorithms} \setlength{\parindent}{0.5cm} + + +\addcontentsline{toc}{chapter}{List of Acronyms} %% Remerciements +\include{ACRONYMS} + %\include{REMERCIEMENTS} %% Citation %\include{CITATION} - - +% LIST OF ACRONYMS \include{Abstruct} diff --git a/Thesis.toc b/Thesis.toc index db993e3..2e06f6f 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -3,103 +3,105 @@ \contentsline {chapter}{List of Figures}{6}{chapter*.2} \contentsline {chapter}{List of Tables}{7}{chapter*.3} \contentsline {chapter}{List of Algorithms}{9}{chapter*.4} -\contentsline {chapter}{Abstract}{11}{chapter*.5} -\contentsline {chapter}{Introduction }{13}{chapter*.6} -\contentsline {section}{1. General Introduction }{13}{section*.7} -\contentsline {section}{2. Motivation of the Dissertation }{14}{section*.8} -\contentsline {section}{3. The Objective of this Dissertation}{14}{section*.9} -\contentsline {section}{4. Main Contributions of this Dissertation}{14}{section*.10} -\contentsline {section}{5. Dissertation Outline}{16}{section*.11} -\contentsline {part}{I\hspace {1em}Scientific Background}{17}{part.1} -\contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{19}{chapter.1} -\contentsline {section}{\numberline {1.1}Introduction}{19}{section.1.1} -\contentsline {section}{\numberline {1.2}Architecture}{20}{section.1.2} -\contentsline {section}{\numberline {1.3}Types of Wireless Sensor Networks}{22}{section.1.3} -\contentsline {section}{\numberline {1.4}Applications}{24}{section.1.4} -\contentsline {section}{\numberline {1.5}The Main Challenges}{27}{section.1.5} -\contentsline {section}{\numberline {1.6}Energy-Efficient Mechanisms of a working WSN}{29}{section.1.6} -\contentsline {subsection}{\numberline {1.6.1}Energy-Efficient Routing}{29}{subsection.1.6.1} -\contentsline {subsubsection}{\numberline {1.6.1.1}Routing Metric based on Residual Energy}{29}{subsubsection.1.6.1.1} -\contentsline {subsubsection}{\numberline {1.6.1.2}Multipath Routing}{29}{subsubsection.1.6.1.2} -\contentsline {subsection}{\numberline {1.6.2}Cluster Architecture}{30}{subsection.1.6.2} -\contentsline {subsection}{\numberline {1.6.3}Scheduling Schemes}{30}{subsection.1.6.3} -\contentsline {subsubsection}{\numberline {1.6.3.1}Wake up Scheduling Schemes}{30}{subsubsection.1.6.3.1} -\contentsline {subsubsection}{\numberline {1.6.3.2}Topology Control Schemes}{33}{subsubsection.1.6.3.2} -\contentsline {subsection}{\numberline {1.6.4}Data-Driven Schemes}{33}{subsection.1.6.4} -\contentsline {subsubsection}{\numberline {1.6.4.1}Data Reduction Schemes}{34}{subsubsection.1.6.4.1} -\contentsline {subsubsection}{\numberline {1.6.4.2}Energy Efficient Data Acquisition Schemes}{34}{subsubsection.1.6.4.2} -\contentsline {subsection}{\numberline {1.6.5}Battery Repletion}{34}{subsection.1.6.5} -\contentsline {subsection}{\numberline {1.6.6}Radio Optimization}{34}{subsection.1.6.6} -\contentsline {subsection}{\numberline {1.6.7}Relay nodes and Sink Mobility}{35}{subsection.1.6.7} -\contentsline {subsubsection}{\numberline {1.6.7.1}Relay node placement}{35}{subsubsection.1.6.7.1} -\contentsline {subsubsection}{\numberline {1.6.7.2}Sink Mobility}{35}{subsubsection.1.6.7.2} -\contentsline {section}{\numberline {1.7}Network Lifetime}{35}{section.1.7} -\contentsline {section}{\numberline {1.8}Coverage in Wireless Sensor Networks }{36}{section.1.8} -\contentsline {section}{\numberline {1.9}Design Issues for Coverage Problems}{38}{section.1.9} -\contentsline {section}{\numberline {1.10}Energy Consumption Model}{39}{section.1.10} -\contentsline {section}{\numberline {1.11}Conclusion}{40}{section.1.11} -\contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{41}{chapter.2} -\contentsline {section}{\numberline {2.1}Introduction}{41}{section.2.1} -\contentsline {section}{\numberline {2.2}Centralized Algorithms}{43}{section.2.2} -\contentsline {section}{\numberline {2.3}Distributed Algorithms}{46}{section.2.3} -\contentsline {subsection}{\numberline {2.3.1}GAF}{48}{subsection.2.3.1} -\contentsline {subsection}{\numberline {2.3.2}DESK}{50}{subsection.2.3.2} -\contentsline {section}{\numberline {2.4}Conclusion}{52}{section.2.4} -\contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{55}{chapter.3} -\contentsline {section}{\numberline {3.1}Introduction}{55}{section.3.1} -\contentsline {section}{\numberline {3.2}Evaluation Tools}{55}{section.3.2} -\contentsline {subsection}{\numberline {3.2.1}Testbed Tools}{56}{subsection.3.2.1} -\contentsline {subsection}{\numberline {3.2.2}Simulation Tools}{57}{subsection.3.2.2} -\contentsline {section}{\numberline {3.3}Optimization Solvers}{62}{section.3.3} -\contentsline {section}{\numberline {3.4}Conclusion}{65}{section.3.4} -\contentsline {part}{II\hspace {1em}Contributions}{67}{part.2} -\contentsline {chapter}{\numberline {4}Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{69}{chapter.4} -\contentsline {section}{\numberline {4.1}Introduction}{69}{section.4.1} -\contentsline {section}{\numberline {4.2}Description of the DiLCO Protocol}{70}{section.4.2} -\contentsline {subsection}{\numberline {4.2.1}Assumptions and Network Model}{70}{subsection.4.2.1} -\contentsline {subsection}{\numberline {4.2.2}Primary Point Coverage Model}{71}{subsection.4.2.2} -\contentsline {subsection}{\numberline {4.2.3}Main Idea}{72}{subsection.4.2.3} -\contentsline {subsubsection}{\numberline {4.2.3.1}Information Exchange Phase}{73}{subsubsection.4.2.3.1} -\contentsline {subsubsection}{\numberline {4.2.3.2}Leader Election Phase}{73}{subsubsection.4.2.3.2} -\contentsline {subsubsection}{\numberline {4.2.3.3}Decision phase}{73}{subsubsection.4.2.3.3} -\contentsline {subsubsection}{\numberline {4.2.3.4}Sensing phase}{73}{subsubsection.4.2.3.4} -\contentsline {section}{\numberline {4.3}Primary Points based Coverage Problem Formulation}{74}{section.4.3} -\contentsline {section}{\numberline {4.4}Simulation Results and Analysis}{76}{section.4.4} -\contentsline {subsection}{\numberline {4.4.1}Simulation Framework}{76}{subsection.4.4.1} -\contentsline {subsection}{\numberline {4.4.2}Modeling Language and Optimization Solver}{76}{subsection.4.4.2} -\contentsline {subsection}{\numberline {4.4.3}Energy Consumption Model}{76}{subsection.4.4.3} -\contentsline {subsection}{\numberline {4.4.4}Performance Metrics}{77}{subsection.4.4.4} -\contentsline {subsection}{\numberline {4.4.5}Performance Analysis for Different Subregions}{78}{subsection.4.4.5} -\contentsline {subsection}{\numberline {4.4.6}Performance Analysis for Primary Point Models}{84}{subsection.4.4.6} -\contentsline {subsection}{\numberline {4.4.7}Performance Comparison with other Approaches}{89}{subsection.4.4.7} -\contentsline {section}{\numberline {4.5}Conclusion}{95}{section.4.5} -\contentsline {chapter}{\numberline {5}Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{97}{chapter.5} -\contentsline {section}{\numberline {5.1}Introduction}{97}{section.5.1} -\contentsline {section}{\numberline {5.2}MuDiLCO Protocol Description}{98}{section.5.2} -\contentsline {subsection}{\numberline {5.2.1}Background Idea and Algorithm}{98}{subsection.5.2.1} -\contentsline {section}{\numberline {5.3}Primary Points based Multiround Coverage Problem Formulation}{99}{section.5.3} -\contentsline {section}{\numberline {5.4}Experimental Study and Analysis}{101}{section.5.4} -\contentsline {subsection}{\numberline {5.4.1}Simulation Setup}{101}{subsection.5.4.1} -\contentsline {subsection}{\numberline {5.4.2}Metrics}{102}{subsection.5.4.2} -\contentsline {subsection}{\numberline {5.4.3}Results Analysis and Comparison }{103}{subsection.5.4.3} -\contentsline {section}{\numberline {5.5}Conclusion}{108}{section.5.5} -\contentsline {chapter}{\numberline {6}Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks}{111}{chapter.6} -\contentsline {section}{\numberline {6.1}Introduction}{111}{section.6.1} -\contentsline {section}{\numberline {6.2}The PeCO Protocol Description}{112}{section.6.2} -\contentsline {subsection}{\numberline {6.2.1}Assumptions and Models}{112}{subsection.6.2.1} -\contentsline {subsection}{\numberline {6.2.2}The Main Idea}{115}{subsection.6.2.2} -\contentsline {subsection}{\numberline {6.2.3}PeCO Protocol Algorithm}{115}{subsection.6.2.3} -\contentsline {section}{\numberline {6.3}Perimeter-based Coverage Problem Formulation}{116}{section.6.3} -\contentsline {section}{\numberline {6.4}Performance Evaluation and Analysis}{118}{section.6.4} -\contentsline {subsection}{\numberline {6.4.1}Simulation Settings}{118}{subsection.6.4.1} -\contentsline {subsection}{\numberline {6.4.2}Simulation Results}{119}{subsection.6.4.2} -\contentsline {subsubsection}{\numberline {6.4.2.1}Coverage Ratio}{120}{subsubsection.6.4.2.1} -\contentsline {subsubsection}{\numberline {6.4.2.2}Active Sensors Ratio}{120}{subsubsection.6.4.2.2} -\contentsline {subsubsection}{\numberline {6.4.2.3}The Energy Consumption}{121}{subsubsection.6.4.2.3} -\contentsline {subsubsection}{\numberline {6.4.2.4}The Network Lifetime}{121}{subsubsection.6.4.2.4} -\contentsline {section}{\numberline {6.5}Conclusion}{124}{section.6.5} -\contentsline {part}{III\hspace {1em}Conclusion and Perspectives}{125}{part.3} -\contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{127}{chapter.7} -\contentsline {section}{\numberline {7.1}Conclusion}{127}{section.7.1} -\contentsline {section}{\numberline {7.2}Perspectives}{128}{section.7.2} -\contentsline {part}{Bibliographie}{144}{chapter*.12} +\contentsline {chapter}{List of Acronyms}{9}{chapter*.4} +\contentsline {chapter}{abbreviations}{11}{chapter*.5} +\contentsline {chapter}{Abstract}{13}{chapter*.6} +\contentsline {chapter}{Introduction }{15}{chapter*.7} +\contentsline {section}{1. General Introduction }{15}{section*.8} +\contentsline {section}{2. Motivation of the Dissertation }{16}{section*.9} +\contentsline {section}{3. The Objective of this Dissertation}{16}{section*.10} +\contentsline {section}{4. Main Contributions of this Dissertation}{16}{section*.11} +\contentsline {section}{5. Dissertation Outline}{18}{section*.12} +\contentsline {part}{I\hspace {1em}Scientific Background}{19}{part.1} +\contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{21}{chapter.1} +\contentsline {section}{\numberline {1.1}Introduction}{21}{section.1.1} +\contentsline {section}{\numberline {1.2}Architecture}{22}{section.1.2} +\contentsline {section}{\numberline {1.3}Types of Wireless Sensor Networks}{24}{section.1.3} +\contentsline {section}{\numberline {1.4}Applications}{26}{section.1.4} +\contentsline {section}{\numberline {1.5}The Main Challenges}{29}{section.1.5} +\contentsline {section}{\numberline {1.6}Energy-Efficient Mechanisms of a working WSN}{31}{section.1.6} +\contentsline {subsection}{\numberline {1.6.1}Energy-Efficient Routing}{31}{subsection.1.6.1} +\contentsline {subsubsection}{\numberline {1.6.1.1}Routing Metric based on Residual Energy}{31}{subsubsection.1.6.1.1} +\contentsline {subsubsection}{\numberline {1.6.1.2}Multipath Routing}{31}{subsubsection.1.6.1.2} +\contentsline {subsection}{\numberline {1.6.2}Cluster Architecture}{32}{subsection.1.6.2} +\contentsline {subsection}{\numberline {1.6.3}Scheduling Schemes}{32}{subsection.1.6.3} +\contentsline {subsubsection}{\numberline {1.6.3.1}Wake up Scheduling Schemes}{32}{subsubsection.1.6.3.1} +\contentsline {subsubsection}{\numberline {1.6.3.2}Topology Control Schemes}{35}{subsubsection.1.6.3.2} +\contentsline {subsection}{\numberline {1.6.4}Data-Driven Schemes}{35}{subsection.1.6.4} +\contentsline {subsubsection}{\numberline {1.6.4.1}Data Reduction Schemes}{36}{subsubsection.1.6.4.1} +\contentsline {subsubsection}{\numberline {1.6.4.2}Energy Efficient Data Acquisition Schemes}{36}{subsubsection.1.6.4.2} +\contentsline {subsection}{\numberline {1.6.5}Battery Repletion}{36}{subsection.1.6.5} +\contentsline {subsection}{\numberline {1.6.6}Radio Optimization}{36}{subsection.1.6.6} +\contentsline {subsection}{\numberline {1.6.7}Relay nodes and Sink Mobility}{37}{subsection.1.6.7} +\contentsline {subsubsection}{\numberline {1.6.7.1}Relay node placement}{37}{subsubsection.1.6.7.1} +\contentsline {subsubsection}{\numberline {1.6.7.2}Sink Mobility}{37}{subsubsection.1.6.7.2} +\contentsline {section}{\numberline {1.7}Network Lifetime}{37}{section.1.7} +\contentsline {section}{\numberline {1.8}Coverage in Wireless Sensor Networks }{38}{section.1.8} +\contentsline {section}{\numberline {1.9}Design Issues for Coverage Problems}{40}{section.1.9} +\contentsline {section}{\numberline {1.10}Energy Consumption Model}{41}{section.1.10} +\contentsline {section}{\numberline {1.11}Conclusion}{42}{section.1.11} +\contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{43}{chapter.2} +\contentsline {section}{\numberline {2.1}Introduction}{43}{section.2.1} +\contentsline {section}{\numberline {2.2}Centralized Algorithms}{45}{section.2.2} +\contentsline {section}{\numberline {2.3}Distributed Algorithms}{48}{section.2.3} +\contentsline {subsection}{\numberline {2.3.1}GAF}{50}{subsection.2.3.1} +\contentsline {subsection}{\numberline {2.3.2}DESK}{52}{subsection.2.3.2} +\contentsline {section}{\numberline {2.4}Conclusion}{54}{section.2.4} +\contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{57}{chapter.3} +\contentsline {section}{\numberline {3.1}Introduction}{57}{section.3.1} +\contentsline {section}{\numberline {3.2}Evaluation Tools}{57}{section.3.2} +\contentsline {subsection}{\numberline {3.2.1}Testbed Tools}{58}{subsection.3.2.1} +\contentsline {subsection}{\numberline {3.2.2}Simulation Tools}{59}{subsection.3.2.2} +\contentsline {section}{\numberline {3.3}Optimization Solvers}{64}{section.3.3} +\contentsline {section}{\numberline {3.4}Conclusion}{67}{section.3.4} +\contentsline {part}{II\hspace {1em}Contributions}{69}{part.2} +\contentsline {chapter}{\numberline {4}Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{71}{chapter.4} +\contentsline {section}{\numberline {4.1}Introduction}{71}{section.4.1} +\contentsline {section}{\numberline {4.2}Description of the DiLCO Protocol}{72}{section.4.2} +\contentsline {subsection}{\numberline {4.2.1}Assumptions and Network Model}{72}{subsection.4.2.1} +\contentsline {subsection}{\numberline {4.2.2}Primary Point Coverage Model}{73}{subsection.4.2.2} +\contentsline {subsection}{\numberline {4.2.3}Main Idea}{74}{subsection.4.2.3} +\contentsline {subsubsection}{\numberline {4.2.3.1}Information Exchange Phase}{75}{subsubsection.4.2.3.1} +\contentsline {subsubsection}{\numberline {4.2.3.2}Leader Election Phase}{75}{subsubsection.4.2.3.2} +\contentsline {subsubsection}{\numberline {4.2.3.3}Decision phase}{75}{subsubsection.4.2.3.3} +\contentsline {subsubsection}{\numberline {4.2.3.4}Sensing phase}{75}{subsubsection.4.2.3.4} +\contentsline {section}{\numberline {4.3}Primary Points based Coverage Problem Formulation}{76}{section.4.3} +\contentsline {section}{\numberline {4.4}Simulation Results and Analysis}{78}{section.4.4} +\contentsline {subsection}{\numberline {4.4.1}Simulation Framework}{78}{subsection.4.4.1} +\contentsline {subsection}{\numberline {4.4.2}Modeling Language and Optimization Solver}{78}{subsection.4.4.2} +\contentsline {subsection}{\numberline {4.4.3}Energy Consumption Model}{78}{subsection.4.4.3} +\contentsline {subsection}{\numberline {4.4.4}Performance Metrics}{79}{subsection.4.4.4} +\contentsline {subsection}{\numberline {4.4.5}Performance Analysis for Different Subregions}{80}{subsection.4.4.5} +\contentsline {subsection}{\numberline {4.4.6}Performance Analysis for Primary Point Models}{86}{subsection.4.4.6} +\contentsline {subsection}{\numberline {4.4.7}Performance Comparison with other Approaches}{91}{subsection.4.4.7} +\contentsline {section}{\numberline {4.5}Conclusion}{97}{section.4.5} +\contentsline {chapter}{\numberline {5}Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{99}{chapter.5} +\contentsline {section}{\numberline {5.1}Introduction}{99}{section.5.1} +\contentsline {section}{\numberline {5.2}MuDiLCO Protocol Description}{100}{section.5.2} +\contentsline {subsection}{\numberline {5.2.1}Background Idea and Algorithm}{100}{subsection.5.2.1} +\contentsline {section}{\numberline {5.3}Primary Points based Multiround Coverage Problem Formulation}{101}{section.5.3} +\contentsline {section}{\numberline {5.4}Experimental Study and Analysis}{103}{section.5.4} +\contentsline {subsection}{\numberline {5.4.1}Simulation Setup}{103}{subsection.5.4.1} +\contentsline {subsection}{\numberline {5.4.2}Metrics}{104}{subsection.5.4.2} +\contentsline {subsection}{\numberline {5.4.3}Results Analysis and Comparison }{105}{subsection.5.4.3} +\contentsline {section}{\numberline {5.5}Conclusion}{110}{section.5.5} +\contentsline {chapter}{\numberline {6}Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks}{113}{chapter.6} +\contentsline {section}{\numberline {6.1}Introduction}{113}{section.6.1} +\contentsline {section}{\numberline {6.2}The PeCO Protocol Description}{114}{section.6.2} +\contentsline {subsection}{\numberline {6.2.1}Assumptions and Models}{114}{subsection.6.2.1} +\contentsline {subsection}{\numberline {6.2.2}The Main Idea}{117}{subsection.6.2.2} +\contentsline {subsection}{\numberline {6.2.3}PeCO Protocol Algorithm}{117}{subsection.6.2.3} +\contentsline {section}{\numberline {6.3}Perimeter-based Coverage Problem Formulation}{118}{section.6.3} +\contentsline {section}{\numberline {6.4}Performance Evaluation and Analysis}{120}{section.6.4} +\contentsline {subsection}{\numberline {6.4.1}Simulation Settings}{120}{subsection.6.4.1} +\contentsline {subsection}{\numberline {6.4.2}Simulation Results}{121}{subsection.6.4.2} +\contentsline {subsubsection}{\numberline {6.4.2.1}Coverage Ratio}{122}{subsubsection.6.4.2.1} +\contentsline {subsubsection}{\numberline {6.4.2.2}Active Sensors Ratio}{122}{subsubsection.6.4.2.2} +\contentsline {subsubsection}{\numberline {6.4.2.3}The Energy Consumption}{123}{subsubsection.6.4.2.3} +\contentsline {subsubsection}{\numberline {6.4.2.4}The Network Lifetime}{123}{subsubsection.6.4.2.4} +\contentsline {section}{\numberline {6.5}Conclusion}{126}{section.6.5} +\contentsline {part}{III\hspace {1em}Conclusion and Perspectives}{127}{part.3} +\contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{129}{chapter.7} +\contentsline {section}{\numberline {7.1}Conclusion}{129}{section.7.1} +\contentsline {section}{\numberline {7.2}Perspectives}{130}{section.7.2} +\contentsline {part}{Bibliographie}{146}{chapter*.13} diff --git a/entete.tex b/entete.tex index c56ba45..7516066 100644 --- a/entete.tex +++ b/entete.tex @@ -18,9 +18,10 @@ %\documentclass[french,book,nopubpage,nodocumentinfo]{spimufcphdthesis} \documentclass[english, book,nopubpage,nodocumentinfo]{spimufcphdthesis} %%-------------------- - - - +\usepackage[acronym,smallcaps]{glossaries} +\newcommand{\abbrlabel}[1]{\makebox[2cm][l]{\textbf{#1}\ \dotfill}} +\newenvironment{abbreviations}{\begin{list}{}{\renewcommand{\makelabel}{\abbrlabel}}}{\end{list}} +\usepackage{tabularx} \usepackage[utf8]{inputenc} \usepackage{enumerate} \usepackage[english]{babel} -- 2.39.5 From 28fe5f530e0e9ae044a4463fd7eb3646cc5dd04c Mon Sep 17 00:00:00 2001 From: ali Date: Mon, 20 Apr 2015 19:01:59 +0200 Subject: [PATCH 02/16] Update by Ali --- ACRONYMS.tex | 57 +++++++++++++- Abstruct.tex | 2 +- CHAPITRE_02.tex | 44 ++++++----- CHAPITRE_03.tex | 10 +-- CHAPITRE_04.tex | 19 ----- CHAPITRE_05.tex | 16 +--- CHAPITRE_06.tex | 25 +----- Thesis.toc | 200 ++++++++++++++++++++++++------------------------ 8 files changed, 186 insertions(+), 187 deletions(-) diff --git a/ACRONYMS.tex b/ACRONYMS.tex index c873dfd..be974d9 100644 --- a/ACRONYMS.tex +++ b/ACRONYMS.tex @@ -14,6 +14,7 @@ \item[DILCO] Distributed Lifetime Coverage Optimization \item[MuDiLCO] Multiround Distributed Lifetime Coverage Optimization \item[PeCO] Perimeter-based Coverage Optimization +\item[OMNeT++] Objective Modular Network Testbed \item[DESK] Distributed Energy-efficient Scheduling for K-coverage \item[GAF] Geographical Adaptive Fidelity \item[PDA] Personal Digital Assistant @@ -28,10 +29,58 @@ \item[MAV] Micro Aerial Vehicle \item[ECG] Electrocardiogram \item[SCADA] Supervisory Control and Data Acquisition -\item[] -\item[] -\item[] -\item[] +\item[QoS] Quality of Service +\item[DSC] Disjoint Set Covers +\item[MIP] Mixed Integer Programming +\item[LP] Linear Programming +\item[GAS] Geometrically based Activity Scheduling +\item[NCG] Node Coverage Grouping +\item[CG] Column Generation +\item[MLP] Maximum-network Lifetime Problem +\item[RMP] Restricted Master Problem +\item[PS] Pricing Subproblem +\item[GRASP] Greedy Randomized Adaptive Search Procedure +\item[VNS] Variable Neighborhood Search +\item[CSB] Cover Sets Balance +\item[CNSC] Correlated Node Set Computing +\item[HREF] High Residual Energy First +\item[SHM] Structural Health Monitoring +\item[ESA] Effective Sensing Area +\item[MSCR] Maximum Sensing Coverage Region +\item[DASSA] Distributed Adaptive Sleep Scheduling Algorithm +\item[DTGA] Distributed Truncated Greedy Algorithm +\item[FIT] Future Internet of the Things +\item[GUI] Graphical User Interface +\item[NED] NEtwork Description +\item[ns-2] Network Simulator-2 +\item[OPNET] Optimized Network Engineering tool +\item[GloMoSim] Global Mobile System Simulator +\item[SENSE] Sensor Network Simulator and Emulator +\item[GTSNetS] Georgia Tech Sensor Network Simulator +\item[GNU] GNU's Not Unix +\item[GLPK] GNU Linear Programming Kit +\item[MPS] Mathematical Programming System +\item[COIN-OR] Linear Programming +\item[BCP] Branch Cut and Price +\item[CBC] COIN-OR Branch and Cut +\item[OPL] Optimization Programming Language +\item[QP] Quadratic Programming +\item[QCP] Quadratically Constrained Programming +\item[MILP] Mixed Integer Linear Programming +\item[MIQP] Mixed-Integer Quadratic Programming +\item[MIQCP] Mixed-Integer Quadratically Constrained Programming +\item[AIMMS] Advanced Interactive Multidimensional Modeling System +\item[AMPL] A Mathematical Programming Language +\item[GAMS] General Algebraic Modeling System +\item[MPL] Mathematical Programming Language +\item[UAV] Unmanned Aerial Vehicle +\item[WSNL] Wireless Sensor Node Leader +\item[MCU] Microcontroller Unit +\item[CR] Coverage Ratio +\item[EC] Energy Consumption +\item[ASR] Active Sensors Ratio +\item[] + \end{abbreviations} \ No newline at end of file diff --git a/Abstruct.tex b/Abstruct.tex index b75522b..6e538ce 100644 --- a/Abstruct.tex +++ b/Abstruct.tex @@ -42,5 +42,5 @@ Extensive simulations are conducted using the discrete event simulator OMNET++ t the WSN lifetime and provides improved coverage performance. -\textbf{KEY WORDS:} Wireless Sensor Networks, Area Coverage, Network Lifetime, Optimization, Scheduling, Distributed Algorithms, Energy-efficiency. +\textbf{KEY WORDS:} Wireless Networks, Wireless Sensor Networks, Area Coverage, Network Lifetime, Optimization, Scheduling, Distributed Algorithms, Centralized Algorithms, Robustness, Connectivity, Parallel Algorithms, Energy-efficiency, Heterogeneous Energy Network, Homogeneous Network. diff --git a/CHAPITRE_02.tex b/CHAPITRE_02.tex index c25e13d..0c823df 100644 --- a/CHAPITRE_02.tex +++ b/CHAPITRE_02.tex @@ -78,9 +78,9 @@ In a distributed algorithms, on the other hand, the decision process is localize \end{table} -In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between each two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication, less processing power for decision, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no a fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to a predefined priority metrics. The local optimal schedule resulted from the optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally optimal solution, so the solution for all the sensing field is near-optimal. +In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between each two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication, less processing power for decision, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no a fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to a predefined priority metrics. The resulted local optimal schedule of optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally optimal solution, so the solution for all the sensing field is near-optimal. -Several algorithms to retain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. Table~\ref{Table1:ch2} summarized the main characteristics of some coverage approaches in previous literatures. In table~\ref{Table1:ch2}, the "SET K-COVER" characteristic refers to the maximum number of disjoint or non-disjoint sets of sensors such that each set cover can assure the coverage for the whole region. The K-COVER algorithm provides a solution with K cover sets in each execution. The k-coverage characteristic refers to that every point inside the monitored area is always covered by at least k active sensors. +Several algorithms to retain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. Table~\ref{Table1:ch2} summarizes the main characteristics of some coverage approaches in previous literatures. In table~\ref{Table1:ch2}, the "SET K-COVER" characteristic refers to the maximum number of disjoint or non-disjoint sets of sensors such that each set cover can assure the coverage for the whole region. The K-COVER algorithm provides a solution with K cover sets in each execution. The k-coverage characteristic refers to that every point inside the monitored area is always covered by at least k active sensors. @@ -88,28 +88,33 @@ Several algorithms to retain the coverage and maximize the network lifetime were \label{ch2:sec:02} The major idea of most centralized algorithms is to divide/organize the sensors into a suitable number of cover sets, where each set completely covers an interest region and to activate these cover sets successively. The centralized algorithms always provide optimal or near-optimal solution since the algorithm has a global view of the whole network. Energy-efficient centralized approaches differ according to several criteria \cite{ref113}, such as the coverage objective (target coverage or area coverage), the node deployment method (random or deterministic), and the heterogeneity of sensor nodes (common sensing range, common battery lifetime). -The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes, which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers such that the number of covers that include an area, summed over all areas, is maximized. +The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes, which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers such that the number of covers that include an area, summed over all areas, is maximized. Their work builds upon previous work in~\cite{ref116} and the generated cover sets do not provide complete coverage of the monitoring zone. -The authors in~\cite{ref115} propose a heuristic to compute the disjoint set covers (DSC). In order to compute the maximum number of covers, they first transform DSC into a maximum-flow problem, which is then formulated as a mixed integer programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$ where -$n$ is the number of sensors and $m$ the number of targets. Compared to algorithm's results of Slijepcevic and Potkonjak \cite{ref116}, their heuristic produces more cover sets with a slight growth rate in execution time. L. Liu et al.~\cite{ref150} formulate the maximum disjoint sets for maintaining target coverage and connectivity problem in WSN. They propose a graph theoretical framework to study the joint problem of connectivity and coverage in a WSN with random deployment of nodes with no restrictions on the sensing and communication ranges of nodes. They propose heuristic algorithms to find the suitable number of nodes to guarantee connectivity and coverage while maximizing network lifetime. +The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$ where +$n$ is the number of sensors and $m$ the number of targets. Compared to algorithm's results of Slijepcevic and Potkonjak \cite{ref116}, their heuristic produces more cover sets with a slight growth rate in execution time. L. Liu et al.~\cite{ref150} formulate the maximum disjoint sets for maintaining target coverage and connectivity problem in WSN. They propose a graph theoretical framework to study the joint problem of connectivity and coverage in a WSN with random deployment of nodes with no restrictions on the sensing and communication ranges of nodes. They propose heuristic algorithms to find the suitable number of nodes to guarantee connectivity and coverage while maximizing network lifetime. %This work did not take into account the sensor node failure, which is an unpredictable event because the two solutions are full centralized algorithms. -Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem to a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find a full coverage sets with virtual radii and transforming the coverage sets to a partial coverage sets by adjusting sensing radii . This framework has four strategies, two of them are designed for the network where the sensors have fixed sensing range and the other two are for the network where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets by the resolution of an integer programming problem. Each cover set is capable of monitoring all the targets of the region of interest. Those covers sets are scheduled periodically. Their algorithm is able to construct the different cover sets in parallel. The results show that their algorithm achieves near-optimal solutions compared to the optimal ones obtained by the exact method. +Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem to a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find full coverage sets with virtual radii and then transform to the coverage sets to a partial coverage sets by adjusting sensing radii . This framework has four strategies, two of them are designed for the network where the sensors have fixed sensing range and the other two are for the network where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{ref229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets, each capable of monitoring all the targets of the region of interest. %Those covers sets are scheduled periodically. +Their algorithm is able to construct the different cover sets in parallel. The results show that their algorithm achieves near-optimal solutions compared to the optimal ones obtained by the resolution of an integer programming. +%exact method. In the case of non-disjoint algorithms~\cite{ref117}, sensors may participate in more than one cover set. In some cases, this may prolong the lifetime of the network in comparison to the disjoint cover set algorithms, but designing algorithms for non-disjoint cover sets generally induces a higher order of complexity. Moreover, in case of a sensor's failure, non-disjoint scheduling policies are less resilient and reliable because a sensor may be involved in more than one cover sets. For instance, Cardei et al.~\cite{ref167} -present a linear programming (LP) solution and a greedy approach to +present a Linear Programming (LP) solution and a greedy approach to extend the sensor network lifetime by organizing the sensors into a maximal number of non-disjoint cover sets. Simulation results show that by allowing sensors to participate in multiple sets, the network -lifetime increases compared with related work~\cite{ref115}. The authors in~\cite{ref148}, address the problem of minimum cost area coverage in which full coverage is performed by using the minimum number of sensors for an arbitrary geometric shape region. A geometric solution to the minimum cost coverage problem under a deterministic deployment is proposed. The probabilistic coverage solution which provides a relationship between the probability of coverage and the number of randomly deployed sensors in an arbitrarily-shaped region is suggested. The authors explained that with a random deployment about seven times more nodes are required to supply full coverage compared to deterministic deployment. The work in~\cite{ref144} address the area coverage problem by proposing a geometrically based activity scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explained that GAS is capable to monitor the target area by using a few sensors as possible and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called node coverage grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other, and the data collected by them in the same group are supposed to be similar. They proved that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. -For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be premature assumption that sensors near to each other sense similar data. The problem of k-coverage over the area of interest in WSNs was addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They checked the connection of the graph via laplacian of the adjacency graph of active sensors in each round. The generation of coverage bitmap by using Minkowski technique, the network is able to providing the desired ratio of coverage. They defined the connected coverage problem as an optimization problem and a centralized genetic algorithm is used to find the solution. +lifetime increases compared with related work~\cite{ref115}. The authors in~\cite{ref148}, address the problem of minimum cost area coverage in which full coverage is performed by using the minimum number of sensors for an arbitrary geometric shape region. A geometric solution to the minimum cost coverage problem under a deterministic deployment is proposed. The probabilistic coverage solution which provides a relationship between the probability of coverage and the number of randomly deployed sensors in an arbitrarily-shaped region is suggested. +%The authors explained that with a random deployment about seven times more nodes are required to supply full coverage compared to deterministic deployment. +The work in~\cite{ref144} address the area coverage problem by proposing a Geometrically based Activity Scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explain that GAS is capable to monitor the target area by using a few sensors as possible and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called Node Coverage Grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other, and the data collected by them in the same group are supposed to be similar. They prove that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. +For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be premature assumption that sensors near to each other sense similar data. The problem of k-coverage over the area of interest in WSNs is addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They check the connection of the graph via laplacian of the adjacency graph of active sensors in each round. %The generation of coverage bitmap by using Minkowski technique, the network is able to providing the desired ratio of coverage. +They define the connected coverage problem as an optimization problem and a centralized genetic algorithm is used to find the solution. -Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSN \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of maximum network lifetime problem (MLP). CG decomposes the problem into a restricted master problem (RMP) and a pricing subproblem (PS). The former maximizes lifetime using an incomplete set of columns, and the latter is used to identify new profitable columns. A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, and second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation and boosted by a greedy randomized adaptive search procedure (GRASP) and a variable neighborhood search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed by sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. +Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSN \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of Maximum-network Lifetime Problem (MLP). CG decomposes the problem into a Restricted Master Problem (RMP) and a Pricing Subproblem (PS). The former maximizes lifetime using an incomplete set of columns, and the latter is used to identify new profitable columns. A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, and second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation and boosted by a Greedy Randomized Adaptive Search Procedure (GRASP) and a Variable Neighborhood Search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. -More recently, the authors in~\cite{ref118}, consider an area coverage optimization algorithm based on linear programming approach to select the minimum number of working sensor nodes, in order to preserve a maximum coverage and to extend the lifetime of the network. The experimental results show that linear programming can provide a fewest number of active nodes and maximize the network lifetime coverage. M. Rebai et al.~\cite{ref141}, formulate the problem of full grid area coverage problem using two integer linear programming models: the first, a model that takes into account only the overall coverage constraint; the second, both the connectivity and the full grid coverage constraints are taken into consideration. This work does not take into account the energy constraint. H. Cheng et al.~\cite{ref119} define a heuristic area coverage algorithm called Cover Sets Balance (CSB), which chooses a set of active nodes using the tuple (data coverage range, residual energy). Then, they introduce a new Correlated Node Set Computing (CNSC) algorithm to find the correlated node set for a given node. After that, they propose a High Residual Energy First (HREF) node selection algorithm to minimize the number of active nodes so as to prolong the network lifetime. X. Liu et al.~\cite{ref143} explain that in some applications of WSNs such as structural health monitoring (SHM) and volcano monitoring, the traditional coverage model which is a geographic area defined for individual sensors is not always valid. For this reason, they define a generalized area coverage model, which is not required to have the coverage area of individual nodes, but only based on a function to determine whether a set of sensor nodes is capable of satisfy the requested monitoring task for a certain area. They propose two approaches for dividing the deployed nodes into suitable cover sets. +More recently, the authors in~\cite{ref118}, consider an area coverage optimization algorithm based on linear programming approach to select the minimum number of working sensor nodes, in order to preserve a maximum coverage and to extend the lifetime of the network. The experimental results show that linear programming can provide a fewest number of active nodes and maximize the network lifetime coverage. M. Rebai et al.~\cite{ref141}, formulate the problem of full grid area coverage problem using two integer linear programming models: the first, a model that takes into account only the overall coverage constraint; the second, both the connectivity and the full grid coverage constraints are taken into consideration. This work does not take into account the energy constraint. H. Cheng et al.~\cite{ref119} define a heuristic area coverage algorithm called Cover Sets Balance (CSB), which chooses a set of active nodes using the tuple (data coverage range, residual energy). Then, they introduce a new Correlated Node Set Computing (CNSC) algorithm to find the correlated node set for a given node. After that, they propose a High Residual Energy First (HREF) node selection algorithm to minimize the number of active nodes so as to prolong the network lifetime. X. Liu et al.~\cite{ref143} explain that in some applications of WSNs such as Structural Health Monitoring (SHM) and volcano monitoring, the traditional coverage model which is a geographic area defined for individual sensors is not always valid. For this reason, they define a generalized area coverage model, which is not required to have the coverage area of individual nodes, but only based on a function to determine whether a set of sensor nodes is capable of satisfy the requested monitoring task for a certain area. They propose two approaches for dividing the deployed nodes into suitable cover sets. @@ -124,16 +129,17 @@ More recently, the authors in~\cite{ref118}, consider an area coverage optimizat Many distributed algorithms have been developed to perform the scheduling so as to preserve coverage, see for example \cite{ref123,ref124,ref125,ref126,ref109,ref127,ref128,ref97}. Localized and distributed algorithms generally result in non-disjoint set covers. -X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision problem, whose goal is to determine whether every point in the area of interest is monitored by at least k sensors. The authors prove that if the perimeters of sensors are sufficiently covered it will be the case for the whole area. They provide an algorithm in $O(nd~log~d)$ time to compute the perimeter coverage of each sensor, where $d$ denotes the maximum number of sensors that are neighboring to a sensor and $n$ is the total number of sensors in the network. Their solutions can be translated to distributed protocols to solve the coverage problem. +X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision problem, whose goal is to determine whether every point in the area of interest is monitored by at least k sensors. The authors prove that if the perimeters of sensors are sufficiently covered it will be the case for the whole area. They provide an algorithm in $O(nd~log~d)$ time to compute the perimeter coverage of each sensor, where $d$ denotes the maximum number of sensors that are neighboring to a sensor and $n$ is the total number of sensors in the network. +%Their solutions can be translated to distributed protocols to solve the coverage problem. Distributed algorithms typically operate in rounds for a predetermined duration. At the beginning of each round, a sensor exchanges information with its neighbors and makes a decision to either remain turned on or to go to sleep for the round. This decision is basically made on simple greedy criteria like the largest uncovered area \cite{ref130} or maximum uncovered targets \cite{ref131}. -Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increase network lifetime via putting in sleep mode some redundant nodes. In this work, the effective sensing area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfill the needed sensing coverage. The authors in~\cite{ref146}, define a maximum sensing coverage region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode, and choosing a smaller number of active sensors so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use the residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. -A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity was proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is capable of build a sparse coverage set in distributed way by means of only connectivity information. This work considers only the communication range of the sensor is smaller two times the sensing range of sensor. Shibo et al.~\cite{ref137} express that the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160}, design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disk of sensor nodes and boundaries. They proposed two mechanisms for the converted target coverage problems to produce cover sets covering the sensing +Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increase network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfill the needed sensing coverage. The authors in~\cite{ref146}, define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode, and choosing a smaller number of active sensors so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use the residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. +A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is capable of build a sparse coverage set in distributed way by means of only connectivity information. This work considers only that the communication range of the sensor is smaller two times the sensing range of sensor. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160}, design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disk of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing field completely. Simulations results show that this approach can prolong the lifetime of the network compared with other works. The works presented in~\cite{ref134,ref135,ref136} focus on coverage-aware, distributed energy-efficient, and distributed clustering methods respectively, which aim at extending the network lifetime, while the coverage is ensured. -In this dissertation, we focused in more detail on two distributed coverage algorithms, GAF and DESK because we compared our proposed coverage optimization protocols with them during performance evaluation. +In this dissertation, we focus in more details on two distributed coverage algorithms, GAF and DESK because we compared our proposed coverage optimization protocols with them during performance evaluation. \subsection{GAF} @@ -181,9 +187,9 @@ in the square grid. \subsection{DESK} \label{ch2:sec:03:2} -The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for k-coverage (DESK), which ensures that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (active or sleep) based on the perimeter coverage model from~\cite{ref133}. +The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for K-coverage (DESK), which ensures that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (active or sleep) based on the perimeter coverage model from~\cite{ref133}. -DESK is based on the result from \cite{ref133}. In \cite{ref133}, the whole area is k-covered if and only if the perimeters of all sensors are k-covered. The coverage level of perimeter of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs, which were posted into the range [0,2$ \pi $]. According to figure~\ref{figp}~(b), the coverage level of sensor $s_i$ can be calculated via traversing the range from 0 to 2$ \pi $. +DESK is based on the result from \cite{ref133}. In \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are k-covered. The coverage level of perimeter of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs, which were posted into the range [0,2$ \pi $]. According to figure~\ref{figp}~(b), the coverage level of sensor $s_i$ can be calculated via traversing the range from 0 to 2$ \pi $. \begin{figure}[h!] @@ -218,7 +224,7 @@ w_{i} = \left \{ \end{equation} Where $\alpha, \beta,$ and $\eta$ are constant, z is a random number between [0; d], where d is a time slot, to avoid the case where two sensors having the same $w_i$ to be active at the same time. $l(e_i, r_i)$ is the function computing the lifetime of sensor $s_i$ in terms of its current remaining energy $e_i$ and its sensing range $r_i$. -DESK uses two types of messages, mACTIVATE message by which a sensor informs others that it becomes active, and mASK2SLEEP by which a sensor suggests a neighbor to go to sleep due to its uselessness. The concept of uselessness or a redundant neighbor refers to one that does not contribute to the perimeter coverage of the considered sensor. This means that the segment of the perimeter of the considered sensor overlapping with that neighbor is already covered by active sensors. +DESK uses two types of messages, mACTIVATE message by which a sensor informs others that it becomes active, and mASK2SLEEP by which a sensor suggests a neighbor to go to sleep due to its uselessness. The concept of uselessness or a redundant neighbor refers to one that does not contribute to the perimeter coverage of the considered sensor. This means that the segment of the perimeter of the considered sensor overlapping with that neighbor is already covered by active sensors. @@ -346,7 +352,7 @@ This chapter describes some coverage proposed problems in the literature, with t The coverage problem is considered as an essential requirement for many applications in WSNs because the better coverage of an area of interest provides better sensing measurements of the physical phenomenon. Therefore, many extensive researches have been focused on that problem. These researches aim at designing mechanisms that efficiently manage or schedule the sensors after deployment, since sensor nodes have a limited battery life. Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead in WSNs. %Whatever the case, this would result in a lower lifetime coverage in WSNs. -As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach that take into account the advantages of both centralized and distributed coverage approaches. This hybrid approaches can provide a good quality coverage and prolong the network lifetime. +As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. This hybrid approaches can provide a good quality coverage and prolong the network lifetime. diff --git a/CHAPITRE_03.tex b/CHAPITRE_03.tex index 4134bc4..a4dd3fa 100644 --- a/CHAPITRE_03.tex +++ b/CHAPITRE_03.tex @@ -54,7 +54,7 @@ The WISEBED~\cite{ref183} is a large-scale WSN testbed with a hierarchical arch \item \textbf{IoT-LAB:} IoT-LAB testbed~\cite{ref184,ref185} supplies a very large scale infrastructure service appropriate for evaluating wireless sensor devices and heterogeneous communicating objects. IoT-LAB includes more than 2700 wireless sensor nodes deployed in six different regions in France. Different kinds of wireless sensor nodes are available, with different processor architectures (MSP430, STM32, and Cortex-A8) and different wireless chips (802.15.4 PHY @ 800 MHz or 2.4 GHz). Sensor nodes are either mobile or fixed and can be used in different topologies throughout all the regions. -IoT-LAB provides web-based reservation and tools for protocols and applications development, along with direct command-line access to the platform. Wireless sensor nodes firmware can be constructed from source and deployed on reserved nodes, application activity can be controlled and observed, power consumption or radio interference can be measured using the offered tools. IoT-LAB is a part of the FIT experimental platform, a set of supplementary elements that enable experimentation with innovative services for academic and industrial users. +IoT-LAB provides web-based reservation and tools for protocols and applications development, along with direct command-line access to the platform. Wireless sensor nodes firmware can be constructed from source and deployed on reserved nodes, application activity can be controlled and observed, power consumption or radio interference can be measured using the offered tools. IoT-LAB is a part of the FIT (Future Internet of the Things) experimental platform, a set of supplementary elements that enable experimentation with innovative services for academic and industrial users. \end{enumerate} @@ -99,7 +99,7 @@ Several simulation tools are available for WSNs, which vary in their characteris \item \textbf{NS2:} -The Network Simulator-2 (ns-2)~\cite{ref191,ref192} is an open source, discrete event, network simulator. The major goal of ns-2 is to provide a simulation environment for wired as well as wireless networks to simulate different protocols with different network topologies. ns-2 is constructed using C++ and the simulation interface is provided via OTcl, an object-oriented dialect of Tcl. The network topology is determined by the users by writing OTcl scripts, and then the main program of ns-2 simulates this topology with fixed parameters. ns-2 provides a graphical view of the network by using network animator (NAM). NAM interface includes control features that allow researchers to forward, pause, stop, and control the simulation. ns-2 is the most common and widely used network simulator for scientific research work. +The Network Simulator-2 (ns-2)~\cite{ref191,ref192} is an open source, discrete event, network simulator. The major goal of ns-2 is to provide a simulation environment for wired as well as wireless networks to simulate different protocols with different network topologies. ns-2 is constructed using C++ and the simulation interface is provided via OTcl, an object-oriented dialect of Tcl. The network topology is determined by the users by writing OTcl scripts, and then the main program of ns-2 simulates this topology with fixed parameters. ns-2 provides a graphical view of the network by using network animator (Nam). Nam interface includes control features that allow researchers to forward, pause, stop, and control the simulation. ns-2 is the most common and widely used network simulator for scientific research work. The next version, ns-3, is considered as a new simulator and a final replacement of ns-2, not a simple extension~\cite{ref194}. The ns-3 project~\cite{ref193} was started in mid-2006 and is still under intensive development. Like ns-2, ns-3 is an open source, discrete-event network simulator targeted essentially for research and educational use~\cite{ref195}. ns-3 supports both simulation and emulation using sockets. It also provides a tracing facility to help users in debugging. @@ -107,15 +107,15 @@ The next version, ns-3, is considered as a new simulator and a final replacement \item \textbf{OMNeT++:} -OMNeT++ (Objective Modular Network Testbed) is an open-source, free, discrete-event, component-based C++ simulation library, modular simulation framework for building network simulators~\cite{ref158,ref203}. Even if OMNeT++ is not a network simulator itself, it is very popular as a network simulation platform for both scientific and industrial communities. The major goal behind the development of OMNeT++ is to provide a strong simulation tool, which can be used by the academic and commercial researchers for simulating different types of networks in a distributed and parallel way~\cite{ref197}. OMNeT++ has an extensive graphical user interface (GUI) and intelligence support. It runs on Windows, Linux, Mac OS~X, and other Unix-like systems, and provides a component architecture for models. Components (modules) are first programmed in C++, then assembled into larger components and models using a high-level language (NED)~\cite{ref198}. Several simulation frameworks can be used with OMNeT++ such as INET, INETMANET, MiXiM, and Castalia, where each of them provides a set of simulation facilities (modelity and soon) and can be used for specific applications. +OMNeT++ (Objective Modular Network Testbed) is an open-source, free, discrete-event, component-based C++ simulation library, modular simulation framework for building network simulators~\cite{ref158,ref203}. Even if OMNeT++ is not a network simulator itself, it is very popular as a network simulation platform for both scientific and industrial communities. The major goal behind the development of OMNeT++ is to provide a strong simulation tool, which can be used by the academic and commercial researchers for simulating different types of networks in a distributed and parallel way~\cite{ref197}. OMNeT++ has an extensive Graphical User Interface (GUI) and intelligence support. It runs on Windows, Linux, Mac OS~X, and other Unix-like systems, and provides a component architecture for models. Components (modules) are first programmed in C++, then assembled into larger components and models using a high-level language (NED)~\cite{ref198}. Several simulation frameworks can be used with OMNeT++ such as INET, INETMANET, MiXiM, and Castalia, where each of them provides a set of simulation facilities (modelity and soon) and can be used for specific applications. \item \textbf{OPNET:} -OPNET (Optimized Network Engineering tool)~\cite{ref192,ref200,ref201} is the first commercial simulation tool developed in 1987 for communication networks. It is a discrete event, object-oriented, general purpose network simulator, which is widely used in industry. It uses C and Java languages. It provides a comprehensive development environment for the specification, simulation, configuration, and performance analysis of the communication network. OPNET allows researchers to develop various models by means of a graphical interface. It provides different types of tools such as Probe Editor, Filter Tool, and Animation Viewer for data collection to model graph and animate the resulting output. Unlike ns-2, OPNET provides modeling for different sensor-specific hardware, such as physical-link transceivers and antennas. It includes sensor-specific models such as ad-hoc connectivity, mobility of nodes, node failure models, modeling of power-consumption, etc. OPNET is, a commercial simulator and the license is very expensive. This represents the main disadvantage of that simulator. +OPNET (Optimized Network Engineering tool)~\cite{ref192,ref200,ref201} is the first commercial simulation tool developed in 1987 for communication networks. It is a discrete event, object-oriented, general purpose network simulator, which is widely used in industry. It uses C and Java languages. It provides a comprehensive development environment for the specification, simulation, configuration, and performance analysis of the communication network. OPNET allows researchers to develop various models by means of a graphical interface. It provides different types of tools such as Probe Editor, Filter Tool, and Animation Viewer for data collection to model graph and animate the resulting output. Unlike ns-2, OPNET provides modeling for different sensor-specific hardware, such as physical-link transceivers and antennas. It includes sensor-specific models such as ad-hoc connectivity, mobility of nodes, node failure models, modeling of power-consumption, etc. OPNET is, a commercial simulator and the license is very expensive. This represents the main disadvantage of that simulator. -\item \textbf{GloMoSim:} +\item \textbf{GloMoSim:} GloMoSim (Global Mobile System Simulator)~\cite{ref202,ref204,ref205} is an open source, well-documented source code and scalable simulation environment developed in 1998 for mobile wireless networks. It uses a library called Parsec, which is an extension of C for parallel programming. The main feature of GloMoSim simulator is the parallel environment. A parallel network simulation is hard due to the communication among the simulated nodes on different machines. Several types of protocols and models are found in GloMoSim including TCP, IEEE 802.11 CSMA/CA, MAC, UDP, HTTP, FTP, CBR, ODMRP, WRP, DSR, MACA, Telnet, AODV, etc. It uses a VT visualization tool to observe and debug these protocols. The GloMoSim tool is designed to be extensible with all protocols implemented as modules in its library. It also uses an object-oriented approach. diff --git a/CHAPITRE_04.tex b/CHAPITRE_04.tex index 8fc2e39..07788db 100644 --- a/CHAPITRE_04.tex +++ b/CHAPITRE_04.tex @@ -8,25 +8,6 @@ \label{ch4} -\iffalse -\section{Summary} -\label{ch4:sec:01} -In this chapter, a Distributed Lifetime Coverage Optimization protocol (DiLCO) to maintain -the coverage and to improve the lifetime in wireless sensor networks is -proposed. The area of interest is first divided into subregions using a -divide-and-conquer method and then the DiLCO protocol is distributed on the -sensor nodes in each subregion. The DiLCO combines two efficient techniques: -leader election for each subregion, followed by an optimization-based planning -of activity scheduling decisions for each subregion. The proposed DiLCO works -into rounds during which a small number of nodes, remaining active for sensing, -is selected to ensure coverage so as to maximize the lifetime of wireless sensor -network. Each round consists of four phases: (i)~Information Exchange, -(ii)~Leader Election, (iii)~Decision, and (iv)~Sensing. The decision process is -carried out by a leader node, which solves an integer program. Compared with -some existing protocols, simulation results show that the proposed protocol can -prolong the network lifetime and improve the coverage performance effectively. - -\fi \section{Introduction} \label{ch4:sec:01} diff --git a/CHAPITRE_05.tex b/CHAPITRE_05.tex index 6bb2bf5..91014fb 100644 --- a/CHAPITRE_05.tex +++ b/CHAPITRE_05.tex @@ -7,19 +7,6 @@ \chapter{Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks} \label{ch5} -\iffalse - -\section{Summary} -\label{ch5:sec:01} -Coverage and lifetime are two paramount problems in Wireless Sensor Networks (WSNs). In this paper, a method called Multiround Distributed Lifetime Coverage -Optimization protocol (MuDiLCO) is proposed to maintain the coverage and to improve the lifetime in wireless sensor networks. The area of interest is first -divided into subregions and then the MuDiLCO protocol is distributed on the sensor nodes in each subregion. The proposed MuDiLCO protocol works in periods -during which sets of sensor nodes are scheduled to remain active for a number of rounds during the sensing phase, to ensure coverage so as to maximize the -lifetime of WSN. The decision process is carried out by a leader node, which solves an integer program to produce the best representative sets to be used -during the rounds of the sensing phase. Compared with some existing protocols, simulation results based on multiple criteria (energy consumption, coverage -ratio, and so on) show that the proposed protocol can prolong efficiently the network lifetime and improve the coverage performance. - -\fi \section{Introduction} \label{ch5:sec:01} @@ -268,8 +255,7 @@ large compared to $W_{\theta}$. \label{ch5:sec:04:01} We conducted a series of simulations to evaluate the efficiency and the relevance of our approach, using the discrete event simulator OMNeT++ -\cite{ref158}. The simulation parameters are summarized in Table~\ref{table3}. Each experiment for a network is run over 25~different random topologies and the results presented hereafter are the average of these -25 runs. +\cite{ref158}. The simulation parameters are summarized in Table~\ref{table3}. Each experiment for a network is run over 25~different random topologies and the results presented hereafter are the average of these 25 runs. %Based on the results of our proposed work in~\cite{idrees2014coverage}, we found as the region of interest are divided into larger subregions as the network lifetime increased. In this simulation, the network are divided into 16 subregions. We performed simulations for five different densities varying from 50 to 250~nodes deployed over a $50 \times 25~m^2 $ sensing field. More diff --git a/CHAPITRE_06.tex b/CHAPITRE_06.tex index 31e696c..bd0cdf8 100644 --- a/CHAPITRE_06.tex +++ b/CHAPITRE_06.tex @@ -7,29 +7,6 @@ \chapter{Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks} \label{ch6} -\iffalse - -\section{Summary} -\label{ch6:sec:01} - -The most important problem in a Wireless Sensor Network (WSN) is to optimize the -use of its limited energy provision so that it can fulfill its monitoring task -as long as possible. Among known available approaches that can be used to -improve power management, lifetime coverage optimization provides activity -scheduling which ensures sensing coverage while minimizing the energy cost. In -this paper, we propose such an approach called Perimeter-based Coverage Optimization -protocol (PeCO). It is a hybrid of centralized and distributed methods: the -region of interest is first subdivided into subregions and our protocol is then -distributed among sensor nodes in each subregion. -The novelty of our approach lies essentially in the formulation of a new -mathematical optimization model based on the perimeter coverage level to schedule -sensors' activities. Extensive simulation experiments have been performed using -OMNeT++, the discrete event simulator, to demonstrate that PeCO can -offer longer lifetime coverage for WSNs in comparison with some other protocols. - - -\fi - \section{Introduction} \label{ch6:sec:01} @@ -412,7 +389,7 @@ be active during at most 20 periods. The values of $\alpha^j_i$ and $\beta^j_i$ have been chosen to ensure a good network coverage and a longer WSN lifetime. We have given a higher priority to -the undercoverage (by setting the $\alpha^j_i$ with a larger value than +the undercoverage (by setting the $\alpha^j_i$ with a larger value than $\beta^j_i$) so as to prevent the non-coverage for the interval~$i$ of the sensor~$j$. On the other hand, we have assigned to $\beta^j_i$ a value which is slightly lower so as to minimize the number of active sensor nodes which contribute diff --git a/Thesis.toc b/Thesis.toc index 2e06f6f..76885fc 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -5,103 +5,103 @@ \contentsline {chapter}{List of Algorithms}{9}{chapter*.4} \contentsline {chapter}{List of Acronyms}{9}{chapter*.4} \contentsline {chapter}{abbreviations}{11}{chapter*.5} -\contentsline {chapter}{Abstract}{13}{chapter*.6} -\contentsline {chapter}{Introduction }{15}{chapter*.7} -\contentsline {section}{1. General Introduction }{15}{section*.8} -\contentsline {section}{2. Motivation of the Dissertation }{16}{section*.9} -\contentsline {section}{3. The Objective of this Dissertation}{16}{section*.10} -\contentsline {section}{4. Main Contributions of this Dissertation}{16}{section*.11} -\contentsline {section}{5. Dissertation Outline}{18}{section*.12} -\contentsline {part}{I\hspace {1em}Scientific Background}{19}{part.1} -\contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{21}{chapter.1} -\contentsline {section}{\numberline {1.1}Introduction}{21}{section.1.1} -\contentsline {section}{\numberline {1.2}Architecture}{22}{section.1.2} -\contentsline {section}{\numberline {1.3}Types of Wireless Sensor Networks}{24}{section.1.3} -\contentsline {section}{\numberline {1.4}Applications}{26}{section.1.4} -\contentsline {section}{\numberline {1.5}The Main Challenges}{29}{section.1.5} -\contentsline {section}{\numberline {1.6}Energy-Efficient Mechanisms of a working WSN}{31}{section.1.6} -\contentsline {subsection}{\numberline {1.6.1}Energy-Efficient Routing}{31}{subsection.1.6.1} -\contentsline {subsubsection}{\numberline {1.6.1.1}Routing Metric based on Residual Energy}{31}{subsubsection.1.6.1.1} -\contentsline {subsubsection}{\numberline {1.6.1.2}Multipath Routing}{31}{subsubsection.1.6.1.2} -\contentsline {subsection}{\numberline {1.6.2}Cluster Architecture}{32}{subsection.1.6.2} -\contentsline {subsection}{\numberline {1.6.3}Scheduling Schemes}{32}{subsection.1.6.3} -\contentsline {subsubsection}{\numberline {1.6.3.1}Wake up Scheduling Schemes}{32}{subsubsection.1.6.3.1} -\contentsline {subsubsection}{\numberline {1.6.3.2}Topology Control Schemes}{35}{subsubsection.1.6.3.2} -\contentsline {subsection}{\numberline {1.6.4}Data-Driven Schemes}{35}{subsection.1.6.4} -\contentsline {subsubsection}{\numberline {1.6.4.1}Data Reduction Schemes}{36}{subsubsection.1.6.4.1} -\contentsline {subsubsection}{\numberline {1.6.4.2}Energy Efficient Data Acquisition Schemes}{36}{subsubsection.1.6.4.2} -\contentsline {subsection}{\numberline {1.6.5}Battery Repletion}{36}{subsection.1.6.5} -\contentsline {subsection}{\numberline {1.6.6}Radio Optimization}{36}{subsection.1.6.6} -\contentsline {subsection}{\numberline {1.6.7}Relay nodes and Sink Mobility}{37}{subsection.1.6.7} -\contentsline {subsubsection}{\numberline {1.6.7.1}Relay node placement}{37}{subsubsection.1.6.7.1} -\contentsline {subsubsection}{\numberline {1.6.7.2}Sink Mobility}{37}{subsubsection.1.6.7.2} -\contentsline {section}{\numberline {1.7}Network Lifetime}{37}{section.1.7} -\contentsline {section}{\numberline {1.8}Coverage in Wireless Sensor Networks }{38}{section.1.8} -\contentsline {section}{\numberline {1.9}Design Issues for Coverage Problems}{40}{section.1.9} -\contentsline {section}{\numberline {1.10}Energy Consumption Model}{41}{section.1.10} -\contentsline {section}{\numberline {1.11}Conclusion}{42}{section.1.11} -\contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{43}{chapter.2} -\contentsline {section}{\numberline {2.1}Introduction}{43}{section.2.1} -\contentsline {section}{\numberline {2.2}Centralized Algorithms}{45}{section.2.2} -\contentsline {section}{\numberline {2.3}Distributed Algorithms}{48}{section.2.3} -\contentsline {subsection}{\numberline {2.3.1}GAF}{50}{subsection.2.3.1} -\contentsline {subsection}{\numberline {2.3.2}DESK}{52}{subsection.2.3.2} -\contentsline {section}{\numberline {2.4}Conclusion}{54}{section.2.4} -\contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{57}{chapter.3} -\contentsline {section}{\numberline {3.1}Introduction}{57}{section.3.1} -\contentsline {section}{\numberline {3.2}Evaluation Tools}{57}{section.3.2} -\contentsline {subsection}{\numberline {3.2.1}Testbed Tools}{58}{subsection.3.2.1} -\contentsline {subsection}{\numberline {3.2.2}Simulation Tools}{59}{subsection.3.2.2} -\contentsline {section}{\numberline {3.3}Optimization Solvers}{64}{section.3.3} -\contentsline {section}{\numberline {3.4}Conclusion}{67}{section.3.4} -\contentsline {part}{II\hspace {1em}Contributions}{69}{part.2} -\contentsline {chapter}{\numberline {4}Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{71}{chapter.4} -\contentsline {section}{\numberline {4.1}Introduction}{71}{section.4.1} -\contentsline {section}{\numberline {4.2}Description of the DiLCO Protocol}{72}{section.4.2} -\contentsline {subsection}{\numberline {4.2.1}Assumptions and Network Model}{72}{subsection.4.2.1} -\contentsline {subsection}{\numberline {4.2.2}Primary Point Coverage Model}{73}{subsection.4.2.2} -\contentsline {subsection}{\numberline {4.2.3}Main Idea}{74}{subsection.4.2.3} -\contentsline {subsubsection}{\numberline {4.2.3.1}Information Exchange Phase}{75}{subsubsection.4.2.3.1} -\contentsline {subsubsection}{\numberline {4.2.3.2}Leader Election Phase}{75}{subsubsection.4.2.3.2} -\contentsline {subsubsection}{\numberline {4.2.3.3}Decision phase}{75}{subsubsection.4.2.3.3} -\contentsline {subsubsection}{\numberline {4.2.3.4}Sensing phase}{75}{subsubsection.4.2.3.4} -\contentsline {section}{\numberline {4.3}Primary Points based Coverage Problem Formulation}{76}{section.4.3} -\contentsline {section}{\numberline {4.4}Simulation Results and Analysis}{78}{section.4.4} -\contentsline {subsection}{\numberline {4.4.1}Simulation Framework}{78}{subsection.4.4.1} -\contentsline {subsection}{\numberline {4.4.2}Modeling Language and Optimization Solver}{78}{subsection.4.4.2} -\contentsline {subsection}{\numberline {4.4.3}Energy Consumption Model}{78}{subsection.4.4.3} -\contentsline {subsection}{\numberline {4.4.4}Performance Metrics}{79}{subsection.4.4.4} -\contentsline {subsection}{\numberline {4.4.5}Performance Analysis for Different Subregions}{80}{subsection.4.4.5} -\contentsline {subsection}{\numberline {4.4.6}Performance Analysis for Primary Point Models}{86}{subsection.4.4.6} -\contentsline {subsection}{\numberline {4.4.7}Performance Comparison with other Approaches}{91}{subsection.4.4.7} -\contentsline {section}{\numberline {4.5}Conclusion}{97}{section.4.5} -\contentsline {chapter}{\numberline {5}Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{99}{chapter.5} -\contentsline {section}{\numberline {5.1}Introduction}{99}{section.5.1} -\contentsline {section}{\numberline {5.2}MuDiLCO Protocol Description}{100}{section.5.2} -\contentsline {subsection}{\numberline {5.2.1}Background Idea and Algorithm}{100}{subsection.5.2.1} -\contentsline {section}{\numberline {5.3}Primary Points based Multiround Coverage Problem Formulation}{101}{section.5.3} -\contentsline {section}{\numberline {5.4}Experimental Study and Analysis}{103}{section.5.4} -\contentsline {subsection}{\numberline {5.4.1}Simulation Setup}{103}{subsection.5.4.1} -\contentsline {subsection}{\numberline {5.4.2}Metrics}{104}{subsection.5.4.2} -\contentsline {subsection}{\numberline {5.4.3}Results Analysis and Comparison }{105}{subsection.5.4.3} -\contentsline {section}{\numberline {5.5}Conclusion}{110}{section.5.5} -\contentsline {chapter}{\numberline {6}Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks}{113}{chapter.6} -\contentsline {section}{\numberline {6.1}Introduction}{113}{section.6.1} -\contentsline {section}{\numberline {6.2}The PeCO Protocol Description}{114}{section.6.2} -\contentsline {subsection}{\numberline {6.2.1}Assumptions and Models}{114}{subsection.6.2.1} -\contentsline {subsection}{\numberline {6.2.2}The Main Idea}{117}{subsection.6.2.2} -\contentsline {subsection}{\numberline {6.2.3}PeCO Protocol Algorithm}{117}{subsection.6.2.3} -\contentsline {section}{\numberline {6.3}Perimeter-based Coverage Problem Formulation}{118}{section.6.3} -\contentsline {section}{\numberline {6.4}Performance Evaluation and Analysis}{120}{section.6.4} -\contentsline {subsection}{\numberline {6.4.1}Simulation Settings}{120}{subsection.6.4.1} -\contentsline {subsection}{\numberline {6.4.2}Simulation Results}{121}{subsection.6.4.2} -\contentsline {subsubsection}{\numberline {6.4.2.1}Coverage Ratio}{122}{subsubsection.6.4.2.1} -\contentsline {subsubsection}{\numberline {6.4.2.2}Active Sensors Ratio}{122}{subsubsection.6.4.2.2} -\contentsline {subsubsection}{\numberline {6.4.2.3}The Energy Consumption}{123}{subsubsection.6.4.2.3} -\contentsline {subsubsection}{\numberline {6.4.2.4}The Network Lifetime}{123}{subsubsection.6.4.2.4} -\contentsline {section}{\numberline {6.5}Conclusion}{126}{section.6.5} -\contentsline {part}{III\hspace {1em}Conclusion and Perspectives}{127}{part.3} -\contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{129}{chapter.7} -\contentsline {section}{\numberline {7.1}Conclusion}{129}{section.7.1} -\contentsline {section}{\numberline {7.2}Perspectives}{130}{section.7.2} -\contentsline {part}{Bibliographie}{146}{chapter*.13} +\contentsline {chapter}{Abstract}{15}{chapter*.6} +\contentsline {chapter}{Introduction }{17}{chapter*.7} +\contentsline {section}{1. General Introduction }{17}{section*.8} +\contentsline {section}{2. Motivation of the Dissertation }{18}{section*.9} +\contentsline {section}{3. The Objective of this Dissertation}{18}{section*.10} +\contentsline {section}{4. Main Contributions of this Dissertation}{18}{section*.11} +\contentsline {section}{5. Dissertation Outline}{20}{section*.12} +\contentsline {part}{I\hspace {1em}Scientific Background}{21}{part.1} +\contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{23}{chapter.1} +\contentsline {section}{\numberline {1.1}Introduction}{23}{section.1.1} +\contentsline {section}{\numberline {1.2}Architecture}{24}{section.1.2} +\contentsline {section}{\numberline {1.3}Types of Wireless Sensor Networks}{26}{section.1.3} +\contentsline {section}{\numberline {1.4}Applications}{28}{section.1.4} +\contentsline {section}{\numberline {1.5}The Main Challenges}{31}{section.1.5} +\contentsline {section}{\numberline {1.6}Energy-Efficient Mechanisms of a working WSN}{33}{section.1.6} +\contentsline {subsection}{\numberline {1.6.1}Energy-Efficient Routing}{33}{subsection.1.6.1} +\contentsline {subsubsection}{\numberline {1.6.1.1}Routing Metric based on Residual Energy}{33}{subsubsection.1.6.1.1} +\contentsline {subsubsection}{\numberline {1.6.1.2}Multipath Routing}{33}{subsubsection.1.6.1.2} +\contentsline {subsection}{\numberline {1.6.2}Cluster Architecture}{34}{subsection.1.6.2} +\contentsline {subsection}{\numberline {1.6.3}Scheduling Schemes}{34}{subsection.1.6.3} +\contentsline {subsubsection}{\numberline {1.6.3.1}Wake up Scheduling Schemes}{34}{subsubsection.1.6.3.1} +\contentsline {subsubsection}{\numberline {1.6.3.2}Topology Control Schemes}{37}{subsubsection.1.6.3.2} +\contentsline {subsection}{\numberline {1.6.4}Data-Driven Schemes}{37}{subsection.1.6.4} +\contentsline {subsubsection}{\numberline {1.6.4.1}Data Reduction Schemes}{38}{subsubsection.1.6.4.1} +\contentsline {subsubsection}{\numberline {1.6.4.2}Energy Efficient Data Acquisition Schemes}{38}{subsubsection.1.6.4.2} +\contentsline {subsection}{\numberline {1.6.5}Battery Repletion}{38}{subsection.1.6.5} +\contentsline {subsection}{\numberline {1.6.6}Radio Optimization}{38}{subsection.1.6.6} +\contentsline {subsection}{\numberline {1.6.7}Relay nodes and Sink Mobility}{39}{subsection.1.6.7} +\contentsline {subsubsection}{\numberline {1.6.7.1}Relay node placement}{39}{subsubsection.1.6.7.1} +\contentsline {subsubsection}{\numberline {1.6.7.2}Sink Mobility}{39}{subsubsection.1.6.7.2} +\contentsline {section}{\numberline {1.7}Network Lifetime}{39}{section.1.7} +\contentsline {section}{\numberline {1.8}Coverage in Wireless Sensor Networks }{40}{section.1.8} +\contentsline {section}{\numberline {1.9}Design Issues for Coverage Problems}{42}{section.1.9} +\contentsline {section}{\numberline {1.10}Energy Consumption Model}{43}{section.1.10} +\contentsline {section}{\numberline {1.11}Conclusion}{44}{section.1.11} +\contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{45}{chapter.2} +\contentsline {section}{\numberline {2.1}Introduction}{45}{section.2.1} +\contentsline {section}{\numberline {2.2}Centralized Algorithms}{47}{section.2.2} +\contentsline {section}{\numberline {2.3}Distributed Algorithms}{50}{section.2.3} +\contentsline {subsection}{\numberline {2.3.1}GAF}{52}{subsection.2.3.1} +\contentsline {subsection}{\numberline {2.3.2}DESK}{53}{subsection.2.3.2} +\contentsline {section}{\numberline {2.4}Conclusion}{56}{section.2.4} +\contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{59}{chapter.3} +\contentsline {section}{\numberline {3.1}Introduction}{59}{section.3.1} +\contentsline {section}{\numberline {3.2}Evaluation Tools}{59}{section.3.2} +\contentsline {subsection}{\numberline {3.2.1}Testbed Tools}{60}{subsection.3.2.1} +\contentsline {subsection}{\numberline {3.2.2}Simulation Tools}{61}{subsection.3.2.2} +\contentsline {section}{\numberline {3.3}Optimization Solvers}{66}{section.3.3} +\contentsline {section}{\numberline {3.4}Conclusion}{69}{section.3.4} +\contentsline {part}{II\hspace {1em}Contributions}{71}{part.2} +\contentsline {chapter}{\numberline {4}Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{73}{chapter.4} +\contentsline {section}{\numberline {4.1}Introduction}{73}{section.4.1} +\contentsline {section}{\numberline {4.2}Description of the DiLCO Protocol}{74}{section.4.2} +\contentsline {subsection}{\numberline {4.2.1}Assumptions and Network Model}{74}{subsection.4.2.1} +\contentsline {subsection}{\numberline {4.2.2}Primary Point Coverage Model}{75}{subsection.4.2.2} +\contentsline {subsection}{\numberline {4.2.3}Main Idea}{76}{subsection.4.2.3} +\contentsline {subsubsection}{\numberline {4.2.3.1}Information Exchange Phase}{77}{subsubsection.4.2.3.1} +\contentsline {subsubsection}{\numberline {4.2.3.2}Leader Election Phase}{77}{subsubsection.4.2.3.2} +\contentsline {subsubsection}{\numberline {4.2.3.3}Decision phase}{77}{subsubsection.4.2.3.3} +\contentsline {subsubsection}{\numberline {4.2.3.4}Sensing phase}{77}{subsubsection.4.2.3.4} +\contentsline {section}{\numberline {4.3}Primary Points based Coverage Problem Formulation}{78}{section.4.3} +\contentsline {section}{\numberline {4.4}Simulation Results and Analysis}{80}{section.4.4} +\contentsline {subsection}{\numberline {4.4.1}Simulation Framework}{80}{subsection.4.4.1} +\contentsline {subsection}{\numberline {4.4.2}Modeling Language and Optimization Solver}{80}{subsection.4.4.2} +\contentsline {subsection}{\numberline {4.4.3}Energy Consumption Model}{80}{subsection.4.4.3} +\contentsline {subsection}{\numberline {4.4.4}Performance Metrics}{81}{subsection.4.4.4} +\contentsline {subsection}{\numberline {4.4.5}Performance Analysis for Different Subregions}{82}{subsection.4.4.5} +\contentsline {subsection}{\numberline {4.4.6}Performance Analysis for Primary Point Models}{88}{subsection.4.4.6} +\contentsline {subsection}{\numberline {4.4.7}Performance Comparison with other Approaches}{93}{subsection.4.4.7} +\contentsline {section}{\numberline {4.5}Conclusion}{99}{section.4.5} +\contentsline {chapter}{\numberline {5}Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{101}{chapter.5} +\contentsline {section}{\numberline {5.1}Introduction}{101}{section.5.1} +\contentsline {section}{\numberline {5.2}MuDiLCO Protocol Description}{102}{section.5.2} +\contentsline {subsection}{\numberline {5.2.1}Background Idea and Algorithm}{102}{subsection.5.2.1} +\contentsline {section}{\numberline {5.3}Primary Points based Multiround Coverage Problem Formulation}{103}{section.5.3} +\contentsline {section}{\numberline {5.4}Experimental Study and Analysis}{105}{section.5.4} +\contentsline {subsection}{\numberline {5.4.1}Simulation Setup}{105}{subsection.5.4.1} +\contentsline {subsection}{\numberline {5.4.2}Metrics}{106}{subsection.5.4.2} +\contentsline {subsection}{\numberline {5.4.3}Results Analysis and Comparison }{107}{subsection.5.4.3} +\contentsline {section}{\numberline {5.5}Conclusion}{112}{section.5.5} +\contentsline {chapter}{\numberline {6}Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks}{115}{chapter.6} +\contentsline {section}{\numberline {6.1}Introduction}{115}{section.6.1} +\contentsline {section}{\numberline {6.2}The PeCO Protocol Description}{116}{section.6.2} +\contentsline {subsection}{\numberline {6.2.1}Assumptions and Models}{116}{subsection.6.2.1} +\contentsline {subsection}{\numberline {6.2.2}The Main Idea}{119}{subsection.6.2.2} +\contentsline {subsection}{\numberline {6.2.3}PeCO Protocol Algorithm}{119}{subsection.6.2.3} +\contentsline {section}{\numberline {6.3}Perimeter-based Coverage Problem Formulation}{120}{section.6.3} +\contentsline {section}{\numberline {6.4}Performance Evaluation and Analysis}{122}{section.6.4} +\contentsline {subsection}{\numberline {6.4.1}Simulation Settings}{122}{subsection.6.4.1} +\contentsline {subsection}{\numberline {6.4.2}Simulation Results}{123}{subsection.6.4.2} +\contentsline {subsubsection}{\numberline {6.4.2.1}Coverage Ratio}{124}{subsubsection.6.4.2.1} +\contentsline {subsubsection}{\numberline {6.4.2.2}Active Sensors Ratio}{124}{subsubsection.6.4.2.2} +\contentsline {subsubsection}{\numberline {6.4.2.3}The Energy Consumption}{125}{subsubsection.6.4.2.3} +\contentsline {subsubsection}{\numberline {6.4.2.4}The Network Lifetime}{125}{subsubsection.6.4.2.4} +\contentsline {section}{\numberline {6.5}Conclusion}{128}{section.6.5} +\contentsline {part}{III\hspace {1em}Conclusion and Perspectives}{129}{part.3} +\contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{131}{chapter.7} +\contentsline {section}{\numberline {7.1}Conclusion}{131}{section.7.1} +\contentsline {section}{\numberline {7.2}Perspectives}{132}{section.7.2} +\contentsline {part}{Bibliographie}{148}{chapter*.13} -- 2.39.5 From 585c6ba024142990d0e8889792b49ca87292c3ee Mon Sep 17 00:00:00 2001 From: ali Date: Fri, 24 Apr 2015 02:00:51 +0200 Subject: [PATCH 03/16] Update by Ali --- ACRONYMS.tex | 3 +-- CHAPITRE_02.tex | 72 +++++++++++++++++++++++++------------------------ CHAPITRE_06.tex | 2 +- Thesis.tex | 6 ++--- Thesis.toc | 13 +++++---- 5 files changed, 48 insertions(+), 48 deletions(-) diff --git a/ACRONYMS.tex b/ACRONYMS.tex index be974d9..f30ab04 100644 --- a/ACRONYMS.tex +++ b/ACRONYMS.tex @@ -1,6 +1,6 @@ \chapter*{abbreviations \markboth{abbreviations}{abbreviations}} \label{chap} -\addcontentsline{toc}{chapter}{abbreviations} +\addcontentsline{toc}{chapter}{List of Abbreviations} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% @@ -79,7 +79,6 @@ \item[CR] Coverage Ratio \item[EC] Energy Consumption \item[ASR] Active Sensors Ratio -\item[] \end{abbreviations} diff --git a/CHAPITRE_02.tex b/CHAPITRE_02.tex index 0c823df..d2a9b87 100644 --- a/CHAPITRE_02.tex +++ b/CHAPITRE_02.tex @@ -9,12 +9,12 @@ \section{Introduction} \label{ch2:sec:01} -The main objective of deploying a large number of wireless sensor nodes in the target area of interest is to construct a WSN, which is responsible for monitoring and observation the sensing field, and detecting the required important event in the area of interest. The coverage problem represents the principle requirement in these applications. The main question that shared by these applications is how can the deployed wireless sensor nodes monitor the physical phenomenon properly. The coverage can be considered as one of the QoS (Quality of Service) parameters, and it is closely related to energy consumption. It represents the sensing task supplied by the wireless sensors in WSNs. +The main objective of deploying a large number of wireless sensor nodes in the target area of interest is to construct a WSN, which is responsible for monitoring the sensing field. The coverage problem represents the principle requirement in these applications. The main question shared by these applications is how can the deployed wireless sensor nodes monitor the physical phenomenon properly. The coverage can be considered as one of the QoS (Quality of Service) parameters, and it is closely related to energy consumption. It represents the sensing task supplied by the wireless sensors in WSNs. -The energy resource limitation of wireless sensor nodes has been considered as a big challenge in order to operate the WSN with less energy consumption whilst fulfill the coverage requirement. The main objective of scattering the wireless sensor nodes over the area of interest is to collect the sensed data of the physical phenomena for processing or reporting, where there are two types of reporting for sensed data in WSNs~\cite{ref138} like event-driven and on-demand. In the latter, the monitoring base station start the reporting operation by transmitting a request to the wireless sensor nodes so as to send their sensed data to the base station; for example, the inventory tracking application. In the former, the reporting operation is triggered by one or more wireless sensor nodes within the physical phenomena by transmitting their sensed data to the controlling base station; for instance, the forest fire detection application. The hybrid scheme of the two types is more flexible. +The energy resource limitation of wireless sensor nodes has been considered as a big challenge. So, it is desired to operate the WSN with less energy consumption whilst fulfilling the coverage requirement. The main objective of scattering the wireless sensor nodes over the area of interest is to collect the sensed data of the physical phenomena for processing or reporting, where there are two types of reporting for sensed data in WSNs~\cite{ref138}: event-driven and on-demand. In the latter, the monitoring base station start the reporting operation by transmitting a request to the wireless sensor nodes so as to send their sensed data to the base station; like in inventory tracking application. In the former, the reporting operation is triggered by one or more wireless sensor nodes within the physical phenomena by transmitting their sensed data to the controlling base station; for instance, the forest fire detection application. In hybrid scheme of the two types is more flexible. -The ultimate goal of the coverage is to ensure that each point in the sensing field is within the sensing range of at least one sensor node. Some applications require high reliability to perform their tasks, so they need that every point in the sensing field is covered by more than one sensor node. In order to avoid the lack in monitoring the area of interest, it is necessary that the WSN are deployed with high density so as to exploit the overlapping among the sensor nodes and to prevent malfunction of sensor nodes in severe environments. The overlap can be exploited by choosing the minimum number of sensor nodes to perform the main tasks of the WSN in the sensing field and putting the rest sensor nodes in very low power sleep mode so as to prolong the network lifetime. This exploitation manner is called sensor activity scheduling that aims to set the activity state of each sensor node in the WSN so that the sensing field can be monitored for a long time as possible. The required level of coverage should be guaranteed by the activity-based scheduling scheme~\cite{ref139}. Many scheduling algorithms have been described in~\cite{ref58,ref57}. +The ultimate goal of the coverage is to ensure that each point in the sensing field is within the sensing range of at least one sensor node. Some applications require high reliability to perform their tasks, so they need that every point in the sensing field is covered by more than one sensor node. In order to avoid a lack of monitoring in the area of interest, it is necessary that WSNs are deployed with high density so as to exploit the overlapping among the sensor nodes and to prevent malfunction of sensor nodes in severe environments. The overlap can be exploited by choosing the minimum number of sensor nodes to perform the main tasks of the WSN in the sensing field and putting the remaining sensor nodes in very low power sleep mode so as to prolong the network lifetime. This exploitation manner, which is called sensor activity scheduling, aims to set the activity state of each sensor node in the WSN so that the sensing field can be monitored for as long as possible. The required level of coverage should be guaranteed by the activity-based scheduling scheme~\cite{ref139}. Many scheduling algorithms have been described in~\cite{ref58,ref57}. %This dissertation focuses on the problem of covering the area of interest as long as possible. Several proposed approaches to extend the network lifetime whilst maintaining the coverage have been viewed in this chapter. M. Cardei and J. Wu~\cite{ref113} have been surveyed the different coverage formulation models and their assumptions, as well as the solutions provided. In~\cite{ref105}, several coverage problems are presented from different angles, where the models and assumptions, as well as proposed solutions in the literatures, are described. In this dissertation, the main contribution of previous works that deal with the coverage problem have been addressed. We end this chapter by focusing on two algorithms, GAF~\cite{GAF} and DESK~\cite{DESK}, since they have been used for comparison against our coverage protocols. @@ -38,17 +38,16 @@ M. Cardei and J. Wu~\cite{ref113} survey the different coverage formulation mode \item Additional requirements for energy-efficient and connected coverage. \end{enumerate} -From our point of view, the choice of non-disjoint or disjoint cover sets (sensors participate or not in many cover sets), coverage type ( area, target, or barrier), coverage ratio, coverage degree (how many sensors are required to cover a target or an area) can be added to the above list. +From our point of view, the choice of non-disjoint or disjoint cover sets (sensors participate or not in many cover sets), coverage type (area, target, or barrier), coverage ratio, coverage degree (how many sensors are required to cover a target or an area) can be added to the above list. -Once a sensor nodes are deployed, a coverage algorithm is run to schedule the sensor nodes into cover sets so as to maintain sufficient coverage in the area of interest and extend the network lifetime. The WSN applications require either complete or partial to area coverage and for target coverage, all the target should be covered. This chapter concentrates only on area coverage and target coverage problems because it is possible to transform the area coverage problem to target ( or point) coverage problem and vice versa. We have excluded the barrier coverage problem from this discussion about the coverage problems because it is outside the scope of this dissertation. -This dissertation focuses mainly on the area coverage problem, where the ultimate goal of the area coverage problem is to choose the minimum number of sensor nodes to cover the whole sensing field. +Once sensor nodes are deployed, a coverage algorithm is run to schedule the sensor nodes into cover sets so as to maintain sufficient coverage in the area of interest and extend the network lifetime. The WSN applications require either complete or partial area coverage, while for target coverage, all the target should be covered. This chapter concentrates only on area coverage and target coverage problems because it is possible to transform the area coverage problem to target (or point) coverage problem and vice versa. We have excluded the barrier coverage problem from this discussion because it is outside the scope of this dissertation. +This dissertation focuses mainly on the area coverage problem, where the ultimate goal is to choose the minimum number of sensor nodes to cover the whole sensing field. %We have focused mainly on the area coverage problem. Therefore, we represent the sensing area of each sensor node in the sensing field as a set of primary points and then achieving full area coverage by covering all the points in the sensing field. The ultimate goal of the area coverage problem is to choose the minimum number of sensor nodes to cover the whole sensing region and prolonging the lifetime of the WSN. -Many centralized and distributed coverage algorithms for activity scheduling have been proposed in the literature and based on different assumptions and objectives. In centralized algorithms, a central controller makes all decisions and distributes the results to sensor nodes. The centralized algorithms have the advantage of requiring very low processing power from the sensor nodes which have usually limited processing capabilities. On the contrary, the exchange of packets in large WSNs may consume a considerable amount of energy in a centralized approach compared to a distributed one. Moreover, centralized approaches usually suffer from the scalability and reliability problems, making them less competitive as the network size increases. - -In a distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and the only information from neighboring nodes are used for the activity decision. Compared to centralized algorithms, distributed algorithms reduce the energy consumption required for radio communication and detection accuracy whilst increase the energy consumption for computation. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give optimal (or near-optimal) solution based only on local information. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. Table~\ref{Table0:ch2} shows a comparison between the centralized coverage algorithms and the distributed coverage algorithms. +Many centralized and distributed coverage algorithms for activity scheduling have been proposed in the literature, based on different assumptions and objectives. In centralized algorithms, a central controller makes all decisions and distributes the results to sensor nodes. The centralized algorithms have the advantage of requiring very low processing power from the sensor nodes which have usually limited processing capabilities. On the contrary, the exchange of packets in large WSNs may consume a considerable amount of energy in a centralized approach compared to a distributed one. Moreover, centralized approaches usually suffer from the scalability and reliability problems, making them less competitive as the network size increases. +In a distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and only informations from neighboring nodes are used for the activity decision. Compared to centralized algorithms, distributed algorithms reduce the energy consumption required for radio communication and detection accuracy whilst the energy consumption for computation is increased. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give an optimal (or near-optimal) solution based only on local informations. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. Table~\ref{Table0:ch2} shows a comparison between centralized coverage algorithms and distributed coverage algorithms. \begin{table}[h] \caption{Centralized Coverage Algorithms vs Distributed Coverage Algorithms} @@ -78,9 +77,12 @@ In a distributed algorithms, on the other hand, the decision process is localize \end{table} -In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between each two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication, less processing power for decision, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no a fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to a predefined priority metrics. The resulted local optimal schedule of optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally optimal solution, so the solution for all the sensing field is near-optimal. +In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication, less energy consumption for processing, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to predefined priority metrics. The resulting local optimal schedule from optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally an optimal solution, so the solution for the whole sensing field is near-optimal. -Several algorithms to retain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. Table~\ref{Table1:ch2} summarizes the main characteristics of some coverage approaches in previous literatures. In table~\ref{Table1:ch2}, the "SET K-COVER" characteristic refers to the maximum number of disjoint or non-disjoint sets of sensors such that each set cover can assure the coverage for the whole region. The K-COVER algorithm provides a solution with K cover sets in each execution. The k-coverage characteristic refers to that every point inside the monitored area is always covered by at least k active sensors. +Several algorithms to maintain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. +Table \ref{x11} summarizes the main characteristics of some coverage approaches in previous literatures. +In this table \ref{x11}, the "SET K-COVER" characteristic refers to the maximum number of disjoint or non-disjoint sets of sensors such that each set cover can assure the coverage for the whole region. +The K-COVER algorithm provides a solution with K cover sets in each execution. The k-coverage characteristic refers to the fact that every point inside the monitored area is always covered by at least k active sensors. @@ -88,19 +90,18 @@ Several algorithms to retain the coverage and maximize the network lifetime were \label{ch2:sec:02} The major idea of most centralized algorithms is to divide/organize the sensors into a suitable number of cover sets, where each set completely covers an interest region and to activate these cover sets successively. The centralized algorithms always provide optimal or near-optimal solution since the algorithm has a global view of the whole network. Energy-efficient centralized approaches differ according to several criteria \cite{ref113}, such as the coverage objective (target coverage or area coverage), the node deployment method (random or deterministic), and the heterogeneity of sensor nodes (common sensing range, common battery lifetime). -The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes, which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers such that the number of covers that include an area, summed over all areas, is maximized. +The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers such that the number of covers that include an area, summed over all areas, is maximized. Their work builds upon previous work in~\cite{ref116} and the generated cover sets do not provide complete coverage of the monitoring zone. -The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$ where +The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC computation into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$, where $n$ is the number of sensors and $m$ the number of targets. Compared to algorithm's results of Slijepcevic and Potkonjak \cite{ref116}, their heuristic produces more cover sets with a slight growth rate in execution time. L. Liu et al.~\cite{ref150} formulate the maximum disjoint sets for maintaining target coverage and connectivity problem in WSN. They propose a graph theoretical framework to study the joint problem of connectivity and coverage in a WSN with random deployment of nodes with no restrictions on the sensing and communication ranges of nodes. They propose heuristic algorithms to find the suitable number of nodes to guarantee connectivity and coverage while maximizing network lifetime. %This work did not take into account the sensor node failure, which is an unpredictable event because the two solutions are full centralized algorithms. -Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem to a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find full coverage sets with virtual radii and then transform to the coverage sets to a partial coverage sets by adjusting sensing radii . This framework has four strategies, two of them are designed for the network where the sensors have fixed sensing range and the other two are for the network where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{ref229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets, each capable of monitoring all the targets of the region of interest. %Those covers sets are scheduled periodically. -Their algorithm is able to construct the different cover sets in parallel. The results show that their algorithm achieves near-optimal solutions compared to the optimal ones obtained by the resolution of an integer programming. +Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem to a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find full coverage sets with virtual radii and then transform the coverage sets to a partial coverage sets by adjusting sensing radii. This framework has four strategies, two of them are designed for network, where the sensors have fixed sensing range and the other two are for network, where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{ref229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets, each capable of monitoring all the targets of the region of interest. %Those covers sets are scheduled periodically. +Their algorithm is able to construct the different cover sets in parallel. The results show that their algorithm achieves near-optimal solutions compared to the optimal ones obtained by the resolution of an integer programming problem. %exact method. - In the case of non-disjoint algorithms~\cite{ref117}, sensors may participate in more than one cover set. In some cases, this may prolong the lifetime of the network in comparison to the disjoint cover set algorithms, but designing algorithms for non-disjoint cover sets generally induces a higher order of complexity. Moreover, in case of a sensor's failure, non-disjoint scheduling policies are less resilient and reliable because a sensor may be involved in more than one cover sets. For instance, Cardei et al.~\cite{ref167} present a Linear Programming (LP) solution and a greedy approach to extend the sensor network lifetime by organizing the sensors into a @@ -108,13 +109,13 @@ maximal number of non-disjoint cover sets. Simulation results show that by allowing sensors to participate in multiple sets, the network lifetime increases compared with related work~\cite{ref115}. The authors in~\cite{ref148}, address the problem of minimum cost area coverage in which full coverage is performed by using the minimum number of sensors for an arbitrary geometric shape region. A geometric solution to the minimum cost coverage problem under a deterministic deployment is proposed. The probabilistic coverage solution which provides a relationship between the probability of coverage and the number of randomly deployed sensors in an arbitrarily-shaped region is suggested. %The authors explained that with a random deployment about seven times more nodes are required to supply full coverage compared to deterministic deployment. -The work in~\cite{ref144} address the area coverage problem by proposing a Geometrically based Activity Scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explain that GAS is capable to monitor the target area by using a few sensors as possible and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called Node Coverage Grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other, and the data collected by them in the same group are supposed to be similar. They prove that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. +The work in~\cite{ref144} address the area coverage problem by proposing a Geometrically based Activity Scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explain that GAS is capable to monitor the target area by using the fewest number of sensors and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called Node Coverage Grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other and the data collected by those in the same group are supposed to be similar. They prove that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be premature assumption that sensors near to each other sense similar data. The problem of k-coverage over the area of interest in WSNs is addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They check the connection of the graph via laplacian of the adjacency graph of active sensors in each round. %The generation of coverage bitmap by using Minkowski technique, the network is able to providing the desired ratio of coverage. They define the connected coverage problem as an optimization problem and a centralized genetic algorithm is used to find the solution. -Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSN \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of Maximum-network Lifetime Problem (MLP). CG decomposes the problem into a Restricted Master Problem (RMP) and a Pricing Subproblem (PS). The former maximizes lifetime using an incomplete set of columns, and the latter is used to identify new profitable columns. A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, and second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation and boosted by a Greedy Randomized Adaptive Search Procedure (GRASP) and a Variable Neighborhood Search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. +Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSNs \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of Maximum-network Lifetime Problem (MLP). CG decomposes the problem into a Restricted Master Problem (RMP) and a Pricing Subproblem (PS). The former maximizes lifetime using an incomplete set of columns and the latter is used to identify new profitable columns. A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation boosted by a Greedy Randomized Adaptive Search Procedure (GRASP) and a Variable Neighborhood Search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. -More recently, the authors in~\cite{ref118}, consider an area coverage optimization algorithm based on linear programming approach to select the minimum number of working sensor nodes, in order to preserve a maximum coverage and to extend the lifetime of the network. The experimental results show that linear programming can provide a fewest number of active nodes and maximize the network lifetime coverage. M. Rebai et al.~\cite{ref141}, formulate the problem of full grid area coverage problem using two integer linear programming models: the first, a model that takes into account only the overall coverage constraint; the second, both the connectivity and the full grid coverage constraints are taken into consideration. This work does not take into account the energy constraint. H. Cheng et al.~\cite{ref119} define a heuristic area coverage algorithm called Cover Sets Balance (CSB), which chooses a set of active nodes using the tuple (data coverage range, residual energy). Then, they introduce a new Correlated Node Set Computing (CNSC) algorithm to find the correlated node set for a given node. After that, they propose a High Residual Energy First (HREF) node selection algorithm to minimize the number of active nodes so as to prolong the network lifetime. X. Liu et al.~\cite{ref143} explain that in some applications of WSNs such as Structural Health Monitoring (SHM) and volcano monitoring, the traditional coverage model which is a geographic area defined for individual sensors is not always valid. For this reason, they define a generalized area coverage model, which is not required to have the coverage area of individual nodes, but only based on a function to determine whether a set of sensor nodes is capable of satisfy the requested monitoring task for a certain area. They propose two approaches for dividing the deployed nodes into suitable cover sets. +More recently, the authors in~\cite{ref118} consider an area coverage optimization algorithm based on linear programming approach to select the minimum number of working sensor nodes, in order to preserve a maximum coverage and to extend the lifetime of the network. The experimental results show that linear programming can provide a fewest number of active nodes and maximize the network lifetime coverage. M. Rebai et al.~\cite{ref141}, formulate the problem of full grid area coverage problem using two integer linear programming models: the first, is in model that takes into account only the overall coverage constraint; the second, both the connectivity and the full grid coverage constraints are taken into consideration. This work does not consider the energy constraint. H. Cheng et al.~\cite{ref119} define a heuristic area coverage algorithm called Cover Sets Balance (CSB), which chooses a set of active nodes using the tuple (data coverage range, residual energy). Then, they introduce a new Correlated Node Set Computing (CNSC) algorithm to find the correlated node set for a given node. After that, they propose a High Residual Energy First (HREF) node selection algorithm to minimize the number of active nodes. X. Liu et al.~\cite{ref143} explain that in some applications of WSNs such as Structural Health Monitoring (SHM) and volcano monitoring, the traditional coverage model which is a geographic area defined for individual sensors is not always valid. For this reason, they define a generalized area coverage model, which is not required to have the coverage area of individual nodes, but only based on a function deciding whether a set of sensor nodes is capable of satisfy the requested monitoring task for a certain area. They propose two approaches for dividing the deployed nodes into suitable cover sets. @@ -129,23 +130,23 @@ More recently, the authors in~\cite{ref118}, consider an area coverage optimizat Many distributed algorithms have been developed to perform the scheduling so as to preserve coverage, see for example \cite{ref123,ref124,ref125,ref126,ref109,ref127,ref128,ref97}. Localized and distributed algorithms generally result in non-disjoint set covers. -X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision problem, whose goal is to determine whether every point in the area of interest is monitored by at least k sensors. The authors prove that if the perimeters of sensors are sufficiently covered it will be the case for the whole area. They provide an algorithm in $O(nd~log~d)$ time to compute the perimeter coverage of each sensor, where $d$ denotes the maximum number of sensors that are neighboring to a sensor and $n$ is the total number of sensors in the network. +X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision problem, whose goal is to determine whether every point in the area of interest is monitored by at least k sensors. The authors prove that if the perimeters of sensors are sufficiently covered it will be the case for the whole area. They provide an algorithm in $O(nd~log~d)$ time to compute the perimeter coverage of each sensor, where $d$ denotes the maximum number of sensors that are neighbors of a sensor and $n$ is the total number of sensors in the network. %Their solutions can be translated to distributed protocols to solve the coverage problem. Distributed algorithms typically operate in rounds for a predetermined duration. At the beginning of each round, a sensor exchanges information with its neighbors and makes a decision to either remain turned on or to go to sleep for the round. This decision is basically made on simple greedy criteria like the largest uncovered area \cite{ref130} or maximum uncovered targets \cite{ref131}. -Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increase network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfill the needed sensing coverage. The authors in~\cite{ref146}, define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode, and choosing a smaller number of active sensors so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use the residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. -A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is capable of build a sparse coverage set in distributed way by means of only connectivity information. This work considers only that the communication range of the sensor is smaller two times the sensing range of sensor. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160}, design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disk of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing +Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increases network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfilling the needed sensing coverage. The authors in~\cite{ref146} define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode. In addition, a smaller number of active sensors is chosen so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. +A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is able to build a sparse coverage set in distributed way by means of only connectivity information. This work considers only that the communication range of the sensor is two times smaller than the sensing one. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160} design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disc of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing field completely. Simulations results show that this approach can prolong the lifetime of the network compared with other works. The works presented in~\cite{ref134,ref135,ref136} focus on coverage-aware, distributed energy-efficient, and distributed clustering methods respectively, which aim at extending the network lifetime, while the coverage is ensured. -In this dissertation, we focus in more details on two distributed coverage algorithms, GAF and DESK because we compared our proposed coverage optimization protocols with them during performance evaluation. +In this dissertation, we focus in more details on two distributed coverage algorithms: GAF and DESK, because we compared our proposed coverage optimization protocols with them during performance evaluation. -\subsection{GAF} +\subsection{Geographical Adaptive Fidelity (GAF)} \label{ch2:sec:03:1} -Xu et al. \cite{GAF} develop an algorithm, called Geographical Adaptive Fidelity (GAF). It uses geographic location information to divide the area of interest into fixed square grids. Within each grid, it keeps only one node staying awake to take the responsibility of sensing and communication. Each sensor node uses its GPS to associate itself with a point in the grid.Figure~\ref{gaf1} gives an example of fixed square grid in GAF. +GAF is developed by Xu et al. \cite{GAF}, it uses geographic location information to divide the area of interest into a fixed square grids. Within each fixed square grid, it keeps only one node staying awake to take the responsibility of sensing and communication. Each sensor node uses its GPS to associate itself with a point in the grid.Figure~\ref{gaf1} gives an example of fixed square grid in GAF. \begin{figure}[h!] \centering @@ -154,13 +155,13 @@ Xu et al. \cite{GAF} develop an algorithm, called Geographical Adaptive Fidelity \label{gaf1} \end{figure} -For two adjacent grids, (for example, A and B in figure~\ref{gaf1}) all sensor nodes inside A can communicate with sensor nodes inside B and vice versa. Therefore, all the sensor nodes are equivalent from the point of view the routing. The size of the fixed grid is based on the radio communication range $R_c$. It is supposed that the fixed grid is square with $r$ units on a side as shown in figure~\ref{gaf1}. The distance between the farthest two possible sensor nodes in two adjacent grids such as, B and C in figure~\ref{gaf1}, should not be greater than the radio communication range $R_c$. For instance, the sensor node \textbf{2} of grid B can communicate with the sensor node \textbf{5} of grid C So, +For two adjacent squares grids, (for example, A and B in figure~\ref{gaf1}) all sensor nodes inside A can communicate with sensor nodes inside B and vice versa. Therefore, all the sensor nodes are equivalent from the point of view the routing. The size of the fixed grid is based on the radio communication range $R_c$. It is supposed that the fixed grid is square with $r$ units on a side as shown in figure~\ref{gaf1}. The distance between the farthest sensor nodes in two adjacent squares, such as B and C in figure~\ref{gaf1}, should not be greater than the radio communication range $R_c$. For instance, the sensor node \textbf{2} of grid B can communicate with the sensor node \textbf{5} of square grid C. Thus, \begin{eqnarray} Distance(2,5) \leq R_c \end{eqnarray} - +and \begin{eqnarray} r^2 + \left(2r \right)^2 \leq R_c^2 \end{eqnarray} @@ -169,8 +170,8 @@ or r \leq \dfrac{R_c}{\sqrt{5}} \end{eqnarray} -The sensor nodes in GAF can be in one of the three states: active, sleep, or discovery. Figure~\ref{gaf2} shows the state transition diagram. Each sensor node is initiated with discovery state. -In discovery state, the radio of each sensor node is turned on. Thereafter, the discovery messages are exchanged among the sensor nodes within the same grid. The discovery message consists of four fields, node id, grid id, estimated node active time (enat), and node state. The node uses its location and grid size to determine the grid id. +The sensor nodes in GAF can be in one of the folling three states: Active, Sleeping, or Discovery. Figure~\ref{gaf2} shows the state transition diagram. Each sensor node is initiated with discovery state. +In discovery state, the radio of each sensor node is turned on. Thereafter, the discovery messages are exchanged among the sensor nodes within the same grid. The discovery message consists of four fields, node id, grid id, estimated node active time (enat), and node state. The node uses its location and grid size to determine the square grid id. \begin{figure}[h!] \centering @@ -187,9 +188,10 @@ in the square grid. \subsection{DESK} \label{ch2:sec:03:2} -The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for K-coverage (DESK), which ensures that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (active or sleep) based on the perimeter coverage model from~\cite{ref133}. +The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for K-coverage (DESK), which ensures that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (Active or Sleep) based on the perimeter coverage model from~\cite{ref133}. -DESK is based on the result from \cite{ref133}. In \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are k-covered. The coverage level of perimeter of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs, which were posted into the range [0,2$ \pi $]. According to figure~\ref{figp}~(b), the coverage level of sensor $s_i$ can be calculated via traversing the range from 0 to 2$ \pi $. +%DESK is based on the result from \cite{ref133}. +In DESK \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are K-covered. The coverage level of perimeter of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs, which were posted into the range [0,2$ \pi $]. According to figure~\ref{figp}~(b), the coverage level of sensor $s_i$ can be calculated via traversing the range from 0 to 2$ \pi $. \begin{figure}[h!] @@ -230,14 +232,15 @@ DESK uses two types of messages, mACTIVATE message by which a sensor informs oth Typically, the algorithm works as follows. At the beginning of each round, no sensors are active. All sensors are in listening mode, i.e. all wait for the time to make a decision while still doing sensing job. All the sensor nodes collect the information (coordinates, current residual energy, and sensing range) from the one-hop neighbors. Each sensor stores this information into a list L in the increasing order of the angle $\alpha $ . Each sensor node set its timer to $w_i$ and initially it is proposed that all of its neighbors need it to join the network. When the sensor node $s_j$ joins the network, it broadcasts a mACTIVATE message to inform all of its 1-hop neighbors about its status change. Its neighbors execute the perimeter coverage model to recalculate its coverage level. If it finds any neighbor u that is useless in covering its perimeter, i.e., the perimeter that u covers is covered by other active neighbors, it will send mASK2SLEEP message to that sensor u. When the sensor node receives mASK2SLEEP message, it updates its counter $n_i$, contribution $c_i$ to coverage level, and recalculate waiting time $w_i$. It then check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e., it receives mASK2SLEEP message from all of its neighbors), then it will send message mGOSLEEP to all of its neighbors telling them that it is about to go to sleep, and set a timer $R_i$ for waking up in next round and at last go to sleep. If the sensor node receives mGOSLEEP message, it removes the neighbor sending that message out of its list L. All the sensors have to decide its status in the decision phase. After that, the active sensors perform the sensing task during the sensing phase. -The period the average +%The period the average -\begin{table} +\begin{table}[h] \begin{flushleft} \centering \caption{Main characteristics of some coverage approaches in previous literatures.} +\label{x11} \begin{tabular}{@{} cl*{13}c @{}} & & \multicolumn{10}{c}{Characteristics} \\[2ex] \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ @@ -337,9 +340,8 @@ The period the average \cmidrule[1pt]{2-14} \end{tabular} \end{flushleft} - -\label{Table1:ch2} + \end{table} diff --git a/CHAPITRE_06.tex b/CHAPITRE_06.tex index bd0cdf8..121ae35 100644 --- a/CHAPITRE_06.tex +++ b/CHAPITRE_06.tex @@ -86,7 +86,7 @@ obtained through the formula: $$\alpha = \arccos \left(\dfrac{Dist(u,v)}{2R_s} Every couple of intersection points is placed on the angular interval $[0,2\pi]$ in a counterclockwise manner, leading to a partitioning of the interval. Figure~\ref{pcm2sensors}(a) illustrates the arcs for the nine neighbors of -sensor $0$ and Figure~\ref{expcm} gives the position of the corresponding arcs +sensor $0$ and Figure~\ref{expcm} gives the position of the corresponding arcs in the interval $[0,2\pi]$. More precisely, we can see that the points are ordered according to the measures of the angles defined by their respective positions. The intersection points are then visited one after another, starting diff --git a/Thesis.tex b/Thesis.tex index e1bfe33..e3e3839 100644 --- a/Thesis.tex +++ b/Thesis.tex @@ -13,7 +13,7 @@ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% - +\include{Abstruct} %% Sommaire \tableofcontents \addcontentsline{toc}{chapter}{Table of Contents} @@ -32,7 +32,7 @@ \setlength{\parindent}{0.5cm} -\addcontentsline{toc}{chapter}{List of Acronyms} +%\addcontentsline{toc}{chapter}{List of Abbreviations} %% Remerciements \include{ACRONYMS} @@ -44,7 +44,7 @@ % LIST OF ACRONYMS -\include{Abstruct} + %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% diff --git a/Thesis.toc b/Thesis.toc index 76885fc..5062f88 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -1,11 +1,10 @@ \select@language {english} -\contentsline {chapter}{Table of Contents}{4}{chapter*.1} -\contentsline {chapter}{List of Figures}{6}{chapter*.2} -\contentsline {chapter}{List of Tables}{7}{chapter*.3} -\contentsline {chapter}{List of Algorithms}{9}{chapter*.4} -\contentsline {chapter}{List of Acronyms}{9}{chapter*.4} -\contentsline {chapter}{abbreviations}{11}{chapter*.5} -\contentsline {chapter}{Abstract}{15}{chapter*.6} +\contentsline {chapter}{Abstract}{1}{chapter*.1} +\contentsline {chapter}{Table of Contents}{6}{chapter*.2} +\contentsline {chapter}{List of Figures}{8}{chapter*.3} +\contentsline {chapter}{List of Tables}{9}{chapter*.4} +\contentsline {chapter}{List of Algorithms}{11}{chapter*.5} +\contentsline {chapter}{List of Abbreviations}{13}{chapter*.6} \contentsline {chapter}{Introduction }{17}{chapter*.7} \contentsline {section}{1. General Introduction }{17}{section*.8} \contentsline {section}{2. Motivation of the Dissertation }{18}{section*.9} -- 2.39.5 From 732c8595b841b3178f4d5180f3c21134af1c8a5a Mon Sep 17 00:00:00 2001 From: ali Date: Fri, 24 Apr 2015 17:12:50 +0200 Subject: [PATCH 04/16] update by ali --- Abstruct.tex | 22 ++++++++-------- CHAPITRE_02.tex | 67 ++++++++++++++++++++++++++----------------------- Thesis.toc | 10 ++++---- 3 files changed, 51 insertions(+), 48 deletions(-) diff --git a/Abstruct.tex b/Abstruct.tex index 6e538ce..103d2ba 100644 --- a/Abstruct.tex +++ b/Abstruct.tex @@ -12,34 +12,32 @@ \emph{ \begin{center} \Large Distributed Coverage Optimization Techniques for Improving Lifetime of Wireless Sensor Networks \end{center}} %\emph{ \begin{center} \large By \end{center}} -\emph{ \begin{center} \large Ali Kadhum Idrees \\ The University of Franche-Comt\'e, 2015 \end{center}} +\emph{ \begin{center} \large Ali Kadhum Idrees \\ University of Franche-Comt\'e, 2015 \end{center}} %\emph{ \begin{center} \large The University of Franche-Comt\'e, 2015 \end{center}} \emph{ \begin{center} \large Supervisors: Raphaël Couturier, Karine Deschinkel, and Michel Salomon \end{center}} -Wireless sensor networks (WSNs) have recently received a great deal of research attention due to their wide range of potential applications. in many different fields. Many important characteristics are provided by the WSNs and the sensors that make them different from other wireless ad-hoc networks, and very promising in a wide range of applications. On the other hand, these characteristics are imposed lots of limitations on the WSNs that would lead to several challenges in the network. These challenges might include coverage, topology control, routing, data fusion, security, and many others. One of the main research challenges faced in wireless sensor networks is to preserve continuously and effectively the coverage of an area of interest to be monitored, while simultaneously preventing as much as possible a network failure due to battery-depleted nodes. +Wireless sensor networks (WSNs) have recently received a great deal of research attention due to their wide range of potential applications. Many important characteristics are provided by the WSNs which make them different from other wireless ad-hoc networks. These characteristics are imposed lots of limitations on the WSNs that would lead to several challenges in the network. These challenges might include coverage, topology control, routing, data fusion, security, and many others. One of the main research challenges faced in wireless sensor networks is to preserve continuously and effectively the coverage of an area of interest to be monitored, while simultaneously preventing as much as possible a network failure due to battery-depleted nodes. -In this dissertation, we highly focus on the area coverage problem, energy-efficiency is also the foremost requirement. We have considered a distributed optimization protocols with the ultimate objective of prolonging the network lifetime. The proposed distributed optimization protocols ( including algorithms, models, and solving integer programs) should be energy-efficient protocols. To address +In this dissertation, we highly focus on the area coverage problem, energy-efficiency is also the foremost requirement. We have considered distributed optimization protocols with the ultimate objective of prolonging the network lifetime. The proposed distributed optimization protocols (including algorithms, models, and solving integer programs) should be energy-efficient protocols. To address this problem, this dissertation proposes two-step approaches. Firstly, the sensing field is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of our proposed distributed optimization protocols is distributed and applied on the -sensor nodes in each subregion so as to optimize the coverage and the lifetime performances. In this dissertation, three coverage optimization protocols are proposed. These protocols combine two efficient techniques: leader election for each subregion, followed by an optimization-based planning of activity scheduling decisions for each subregion. +sensor nodes in each subregion so as to optimize the coverage and the lifetime performances. In this dissertation, three coverage optimization protocols are proposed. These protocols combine two efficient techniques: leader election for each subregion, followed by an optimization-based planning of sensor activity scheduling decisions for each subregion. -First, we propose a Distributed Lifetime Coverage Optimization (DILCO) protocol in WSNs. In this protocol, the lifetime is divided into periods. Each period consists of 4 phases: information exchange, leader election, decision, and sensing. The decision process is +First, we propose a protocol called Distributed Lifetime Coverage Optimization (DILCO). In this protocol, the lifetime is divided into periods. Each period consists of 4 phases: information exchange, leader election, decision, and sensing. The decision process is carried out by a leader node, which solves an integer program in order to provide only one cover set of active sensor nodes to ensure coverage during the sensing phase of the current period. - Then we move to address the problem of a multiround optimization of the area coverage problem in WSNs. The Multiround Distributed Lifetime Coverage Optimization (MuDiLCO) protocol is suggested so as to study the possibility of providing multiple cover sets of sensors for the sensing phase. MuDiLCO protocol also works in periods -during which sets of sensor nodes are scheduled to remain active for a number of rounds during the sensing phase, to ensure coverage so as to maximize the -lifetime of WSN. The decision process is carried out by a leader node, which solves an integer program to produce the best representative sets to be used during the rounds of the sensing phase. + Then we address the problem of a multiround optimization of the area coverage problem in WSNs. The Multiround Distributed Lifetime Coverage Optimization (MuDiLCO) protocol is suggested so as to study the possibility of providing multiple cover sets of sensors for the sensing phase. MuDiLCO protocol also works in periods +during which sets of sensor nodes are scheduled to remain active for a number of rounds during the sensing phase, to ensure coverage so as to maximize the lifetime of WSN. The decision process is still carried out by a leader node, which solves an integer program to produce the best representative sets to be used during the rounds of the sensing phase. -Last but not least, we propose a Perimeter-based Coverage Optimization (PeCO) protocol to maintain the coverage and improve the network lifetime in WSNs. It is a hybrid of centralized and distributed methods, where it is also distributed among sensor nodes in each subregion.The novelty of our approach lies essentially in the formulation of a new -mathematical optimization model based on the perimeter coverage level to schedule -sensors' activities. The PeCO protocol resolves a new integer program coverage model by the leader during the decision phase so as to provide only one cover set of sensors for the sensing phase. +Last but not least, we propose a Perimeter-based Coverage Optimization (PeCO) protocol which is also distributed among sensor nodes in each subregion.The novelty of our approach lies essentially in the formulation of a new +mathematical optimization model based on the perimeter coverage level to schedule sensors' activities. A new integer program coverage model is solved by the leader during the decision phase so as to provide only one cover set of sensors for the sensing phase. Extensive simulations are conducted using the discrete event simulator OMNET++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase -the WSN lifetime and provides improved coverage performance. +the WSN lifetime and provide improved coverage performance. \textbf{KEY WORDS:} Wireless Networks, Wireless Sensor Networks, Area Coverage, Network Lifetime, Optimization, Scheduling, Distributed Algorithms, Centralized Algorithms, Robustness, Connectivity, Parallel Algorithms, Energy-efficiency, Heterogeneous Energy Network, Homogeneous Network. diff --git a/CHAPITRE_02.tex b/CHAPITRE_02.tex index d2a9b87..d13ee68 100644 --- a/CHAPITRE_02.tex +++ b/CHAPITRE_02.tex @@ -49,7 +49,7 @@ Many centralized and distributed coverage algorithms for activity scheduling hav In a distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and only informations from neighboring nodes are used for the activity decision. Compared to centralized algorithms, distributed algorithms reduce the energy consumption required for radio communication and detection accuracy whilst the energy consumption for computation is increased. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give an optimal (or near-optimal) solution based only on local informations. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. Table~\ref{Table0:ch2} shows a comparison between centralized coverage algorithms and distributed coverage algorithms. -\begin{table}[h] +\begin{table}[h!] \caption{Centralized Coverage Algorithms vs Distributed Coverage Algorithms} \begin{center} \begin{tabular}{ |p{3cm}|p{5cm}|p{5cm}|} @@ -57,7 +57,7 @@ In a distributed algorithms, on the other hand, the decision process is localize \textbf{\begin{center} Characteristics \end{center}} & \textbf{\begin{center} Centralized Coverage Algorithms \end{center}} & \textbf{\begin{center} Distributed Coverage Algorithms \end{center}}\\ \hline -\textbf{\begin{center} Computation \end{center}} & Require low processing power where the algorithm is executed only in one elected node. & Require large processing power due to execution the algorithm in every node in WSN. \\ \hline +\textbf{\begin{center} Computation \end{center}} & Require low processing power where the algorithm is executed only in one node. & Require large processing power due to executing the algorithm in every node in WSN. \\ \hline \textbf{\begin{center} Communication \end{center}} & Sensor nodes communicate directly with the base station, therefore, they require low-power consumption for communication. & Sensor nodes require high power consumption for communication because of the frequent exchange of hello packets. \\ \hline @@ -67,7 +67,7 @@ In a distributed algorithms, on the other hand, the decision process is localize \textbf{\begin{center} Energy Consumption \end{center}} & Energy consumption is large especially when the network size and/or density increase. & Energy consumption is low because they have lower communication cost. \\ \hline -\textbf{\begin{center} Scalability \end{center}} & Scalable only with dividing the sensing field into smaller subregions. & More scalable for large networks. \\ \hline +\textbf{\begin{center} Scalability \end{center}} & Not scalable, but they can overcome this problem by dividing the sensing field into smaller subregions. & More scalable for large networks. \\ \hline \textbf{\begin{center} Reliability \end{center}} & Less robust against sensor failure. & More robust against sensor failure. \\ \hline @@ -77,7 +77,7 @@ In a distributed algorithms, on the other hand, the decision process is localize \end{table} -In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication, less energy consumption for processing, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to predefined priority metrics. The resulting local optimal schedule from optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally an optimal solution, so the solution for the whole sensing field is near-optimal. +In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication and processing, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to predefined priority metrics. The resulting local optimal schedule from optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally an optimal solution, so the solution for the whole sensing field is near-optimal. Several algorithms to maintain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. Table \ref{x11} summarizes the main characteristics of some coverage approaches in previous literatures. @@ -91,10 +91,10 @@ The K-COVER algorithm provides a solution with K cover sets in each execution. T The major idea of most centralized algorithms is to divide/organize the sensors into a suitable number of cover sets, where each set completely covers an interest region and to activate these cover sets successively. The centralized algorithms always provide optimal or near-optimal solution since the algorithm has a global view of the whole network. Energy-efficient centralized approaches differ according to several criteria \cite{ref113}, such as the coverage objective (target coverage or area coverage), the node deployment method (random or deterministic), and the heterogeneity of sensor nodes (common sensing range, common battery lifetime). The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers such that the number of covers that include an area, summed over all areas, is maximized. -Their work builds upon previous work in~\cite{ref116} and the generated cover sets do not provide complete coverage of the monitoring zone. +Their work is built upon previous work in~\cite{ref116} and the generated cover sets do not provide complete coverage of the monitoring zone. -The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC computation into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$, where +The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$, where $n$ is the number of sensors and $m$ the number of targets. Compared to algorithm's results of Slijepcevic and Potkonjak \cite{ref116}, their heuristic produces more cover sets with a slight growth rate in execution time. L. Liu et al.~\cite{ref150} formulate the maximum disjoint sets for maintaining target coverage and connectivity problem in WSN. They propose a graph theoretical framework to study the joint problem of connectivity and coverage in a WSN with random deployment of nodes with no restrictions on the sensing and communication ranges of nodes. They propose heuristic algorithms to find the suitable number of nodes to guarantee connectivity and coverage while maximizing network lifetime. %This work did not take into account the sensor node failure, which is an unpredictable event because the two solutions are full centralized algorithms. Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem to a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find full coverage sets with virtual radii and then transform the coverage sets to a partial coverage sets by adjusting sensing radii. This framework has four strategies, two of them are designed for network, where the sensors have fixed sensing range and the other two are for network, where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{ref229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets, each capable of monitoring all the targets of the region of interest. %Those covers sets are scheduled periodically. @@ -110,7 +110,7 @@ that by allowing sensors to participate in multiple sets, the network lifetime increases compared with related work~\cite{ref115}. The authors in~\cite{ref148}, address the problem of minimum cost area coverage in which full coverage is performed by using the minimum number of sensors for an arbitrary geometric shape region. A geometric solution to the minimum cost coverage problem under a deterministic deployment is proposed. The probabilistic coverage solution which provides a relationship between the probability of coverage and the number of randomly deployed sensors in an arbitrarily-shaped region is suggested. %The authors explained that with a random deployment about seven times more nodes are required to supply full coverage compared to deterministic deployment. The work in~\cite{ref144} address the area coverage problem by proposing a Geometrically based Activity Scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explain that GAS is capable to monitor the target area by using the fewest number of sensors and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called Node Coverage Grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other and the data collected by those in the same group are supposed to be similar. They prove that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. -For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be premature assumption that sensors near to each other sense similar data. The problem of k-coverage over the area of interest in WSNs is addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They check the connection of the graph via laplacian of the adjacency graph of active sensors in each round. %The generation of coverage bitmap by using Minkowski technique, the network is able to providing the desired ratio of coverage. +For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be a premature assumption that sensors near to each other sense similar data. The problem of k-coverage over the area of interest in WSNs is addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They check the connection of the graph via laplacian of the adjacency graph of active sensors in each round. %The generation of coverage bitmap by using Minkowski technique, the network is able to providing the desired ratio of coverage. They define the connected coverage problem as an optimization problem and a centralized genetic algorithm is used to find the solution. Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSNs \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of Maximum-network Lifetime Problem (MLP). CG decomposes the problem into a Restricted Master Problem (RMP) and a Pricing Subproblem (PS). The former maximizes lifetime using an incomplete set of columns and the latter is used to identify new profitable columns. A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation boosted by a Greedy Randomized Adaptive Search Procedure (GRASP) and a Variable Neighborhood Search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. @@ -176,24 +176,28 @@ In discovery state, the radio of each sensor node is turned on. Thereafter, the \begin{figure}[h!] \centering \includegraphics[scale=0.4]{Figures/ch2/GAF2.eps} -\caption{ Example of fixed square grid in GAF.} +\caption{ State transitions in GAF.} \label{gaf2} \end{figure} -The sensor node sets a timer to $T_d$ seconds after entering in the discovery state. As soon as the timer fires, the sensor node broadcasts its discovery message and enters the active state. The active sensor node sets a timeout value $T_a$ to define how long it can stay in the active state. After $T_a$, the sensor node will return to the discovery state. Whilst, during its active state, it re-broadcasts its discovery message at intervals $T_d$ periodically. The sensor node with discovery or active state can change its state to sleep when it detects that some other equivalent node will handle routing inside the grid. The sensor nodes in the sleeping state wake up after a sleeping time $T_s$ and go back to the discovery state. In GAF, load balancing is performed by means of periodic election of the leader (i.e., the active node that handle the routing inside the fixed grid). Inside each fixed square grid, sensor nodes collaborate with each other to play different roles. For example, nodes will elect -one sensor node (based on the remaining energy of sensor nodes inside the fixed square grid) to stay awake for a certain period of time, and then the rest go to sleep. This sensor node is responsible for monitoring and reporting data to the base station on behalf of the nodes -in the square grid. +The sensor node sets a timer to $T_d$ seconds after entering in the discovery state. As soon as the timer fires, the sensor node broadcasts its discovery message and enters the active state. The active sensor node sets a timeout value $T_a$ to define how long it can stay in the active state. After $T_a$, the sensor node will return to the discovery state. Sensor node changes its state to Discovery to give a chance to other nodes within the same grid to become Active. +%Whilst, during its active state, it re-broadcasts its discovery message at intervals $T_d$ periodically. +The sensor node with Discovery or Active state can change its state to Sleeping when it detects that some other equivalent node will handle routing inside the grid. The sensor nodes in the Sleeping state wake up after a sleeping time $T_s$ and go back to the Discovery state. In GAF, load balancing is performed by means of periodic election of the leader (i.e., the active node that handle the routing inside the fixed grid). Inside each fixed square grid, sensor nodes collaborate with each other to play different roles. For example, nodes will elect +one sensor node (based on the remaining energy of sensor nodes inside the fixed square grid) to stay awake for a certain period of time, and then the rest go to sleep. This sensor node is responsible for monitoring, routing, and reporting data to the base station on behalf of the nodes in the square grid. For nodes with same state, GAF gives nodes with longer expected lifetime (enat) higher rank, therefore they are called high rank nodes. %A rank-based election algorithm has been used to elect the leader. It is based on the remaining energy of sensor nodes inside the fixed square grid so as to extend the network lifetime. -\subsection{DESK} +\subsection{Distributed Energy-efficient Scheduling for K-coverage (DESK)} \label{ch2:sec:03:2} -The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for K-coverage (DESK), which ensures that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (Active or Sleep) based on the perimeter coverage model from~\cite{ref133}. +% The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for K-coverage (DESK), which +DESK is a novel distributed heuristic to ensure that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied~\cite{DESK}. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (Active or Sleep) based on the perimeter coverage model from~\cite{ref133}. %DESK is based on the result from \cite{ref133}. -In DESK \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are K-covered. The coverage level of perimeter of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs, which were posted into the range [0,2$ \pi $]. According to figure~\ref{figp}~(b), the coverage level of sensor $s_i$ can be calculated via traversing the range from 0 to 2$ \pi $. - +In DESK \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are K-covered. The coverage level of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs in the range [0,2$ \pi $]. According to figure~\ref{figp}~(a) and (b), the coverage level of sensor $s_i$ can be calculated as follows. +%via traversing the range from 0 to 2$ \pi $. +For each sensor $s_j$ such that $d(s_i,s_j)$ $<$ $2R_s$, calculate the angle of $s_i$'s arc, denoted by [$\alpha_{j,L}$, $\alpha_{j,R}$], which is perimeter covered by $s_j$, where $\alpha= arccos(d(s_i, s_j)/2R_s)$ and $d(s_i,s_j)$ is the Euclidean distance between $s_i$ and $s_j$. After that, locate the points $\alpha_{j,L}$ and $\alpha_{j,R}$ of each neighboring sensor $s_j$ of $s_i$ on the line segment $[0, 2\pi]$. These points are sorted in ascending order into a list L. Traverse the line segment from 0 to $2\pi$ by visiting each element in the sorted list L from the left to the right and determine the perimeter coverage of $s_i$. Whenever an element $\alpha_{j,L}$ is traversed, the level of perimeter coverage should be increased by one. Whenever an element $\alpha_{j,R}$ is traversed, the level of perimeter coverage should be decreased by one. + \begin{figure}[h!] \centering \begin{tabular}{@{}cr@{}} @@ -213,25 +217,25 @@ In DESK \cite{ref133}, the whole area is K-covered if and only if the perimeters \label{desk} \end{figure} -Figure~\ref{desk} shows the DESK network time line. DESK works into rounds fashion. The network lifetime is divided into R rounds. Each round consists of two phases: decision phase and sensing phase. The length of round is dRound that means each sensor node executes this algorithm every dRound unit of time. The decision phase at the starting of each round should be taken within W unit of time, where $W<< dRound$ as shown in figure~\ref{desk}. All the sensor nodes should be temporarily awakened in the decision phase so as to decide its status. Every sensor node $s_i$ decides its status to be active or sleep after $w_i$ of waiting time. The waiting time $w_i$ is dynamic and it can be changed at any time based on the status of its sensor neighbors, the remaining energy $e_i$ of $s_i$, and its contribution $c_i$ in the coverage level of the network, where $c_i$ is defined as the number of the neighbors $n_i$ who need $s_i$ to be active. The waiting time is defined as follow +Figure~\ref{desk} shows the DESK network time line. DESK works into rounds fashion. The network lifetime is divided into R rounds. Each round consists of two phases: decision phase and sensing phase. The length of round is dRound that means each sensor node executes this algorithm every dRound unit of time. The decision should be taken within W unit of time, where $W<< dRound$ as shown in figure~\ref{desk}. All the sensor nodes should be temporarily awakened in the decision phase so as to decide their status. Every sensor node $s_i$ decides its status to be active or sleep after $w_i$ of waiting time. The waiting time $w_i$ of node $s_i$ is dynamic and can be changed at any time based on the status of its neighbors, the remaining energy $e_i$ of $s_i$, and its contribution $c_i$ in the coverage level of the network, where $c_i$ is defined as the number of neighbors which need $s_i$ to be active. The waiting time is defined as follows: \begin{equation} w_{i} = \left \{ \begin{array}{ll} - \dfrac{\eta}{n_i^\alpha l(e_i,r_i)^\beta} * W + z & \mbox{If $e_i \geq e_{threshold}$} \\ - W & \mbox{Otherwise.}\\ + \dfrac{\eta}{n_i^\alpha l(e_i,r_i)^\beta} * W + z & \mbox{if $e_i \geq e_{threshold}$} \\ + W & \mbox{otherwise,}\\ \end{array} \right. %\label{eq12} \notag \end{equation} -Where $\alpha, \beta,$ and $\eta$ are constant, z is a random number between [0; d], where d is a time slot, to avoid the case where two sensors having the same $w_i$ to be active at the same time. $l(e_i, r_i)$ is the function computing the lifetime of sensor $s_i$ in terms of its current remaining energy $e_i$ and its sensing range $r_i$. -DESK uses two types of messages, mACTIVATE message by which a sensor informs others that it becomes active, and mASK2SLEEP by which a sensor suggests a neighbor to go to sleep due to its uselessness. The concept of uselessness or a redundant neighbor refers to one that does not contribute to the perimeter coverage of the considered sensor. This means that the segment of the perimeter of the considered sensor overlapping with that neighbor is already covered by active sensors. +where $\alpha, \beta,$ and $\eta$ are constant, z is a random number between [0; d], where d is a time duration, to avoid the case where two sensors to be active at the same time. $l(e_i, r_i)$ is the function computing the lifetime of sensor $s_i$ in terms of its current remaining energy $e_i$ and its sensing range $r_i$. +DESK uses two types of messages, mACTIVATE message by which a sensor informs others that it becomes active, and mASK2SLEEP by which a sensor suggests a neighbor to go to sleep due to its uselessness. The concept of uselessness (or redundancy) of a neighbor refers to one that does not contribute to the perimeter coverage of the considered sensor. This means that the segment of the perimeter of the considered sensor overlapping with that neighbor is already covered by active sensors. -Typically, the algorithm works as follows. At the beginning of each round, no sensors are active. All sensors are in listening mode, i.e. all wait for the time to make a decision while still doing sensing job. All the sensor nodes collect the information (coordinates, current residual energy, and sensing range) from the one-hop neighbors. Each sensor stores this information into a list L in the increasing order of the angle $\alpha $ . Each sensor node set its timer to $w_i$ and initially it is proposed that all of its neighbors need it to join the network. When the sensor node $s_j$ joins the network, it broadcasts a mACTIVATE message to inform all of its 1-hop neighbors about its status change. Its neighbors execute the perimeter coverage model to recalculate its coverage level. If it finds any neighbor u that is useless in covering its perimeter, i.e., the perimeter that u covers is covered by other active neighbors, it will send mASK2SLEEP message to that sensor u. When the sensor node receives mASK2SLEEP message, it updates its counter $n_i$, contribution $c_i$ to coverage level, and recalculate waiting time $w_i$. It then -check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e., it receives mASK2SLEEP message from all of its neighbors), then it will send message mGOSLEEP to all of its neighbors telling them that it is about to go to sleep, and set a timer $R_i$ for waking up in next round and at last go to sleep. If the sensor node receives mGOSLEEP message, it removes the neighbor sending that message out of its list L. All the sensors have to decide its status in the decision phase. After that, the active sensors perform the sensing task during the sensing phase. +Typically, the algorithm works as follows. At the beginning of each round, there are no active sensors. All sensors are in listening mode, i.e. all wait for the time to make a decision while still doing sensing job. All the sensor nodes collect the information (coordinates, current residual energy, and sensing range) from the one-hop neighbors. Each sensor stores this information into a list L in the increasing order of the angle $\alpha $. Each sensor sets its timer $w_i$ with the assumption that all of its neighbors need it to join the network. When the sensor node $s_j$ joins the network, it broadcasts a mACTIVATE message to inform all of its one hop neighbors about its status change. Its neighbors execute the perimeter coverage model to recalculate their coverage level. If a node finds any neighbor u that is useless in covering its perimeter, i.e., the perimeter that u covers is covered by other active neighbors, it will send mASK2SLEEP message to that sensor u. When the sensor node receives mASK2SLEEP message, it updates its counter $n_i$, contribution $c_i$ to coverage level, and recalculate waiting time $w_i$. It then +check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e., it receives mASK2SLEEP message from all of its neighbors), then it will send message mGOSLEEP to all of its neighbors telling them that it is about to go to sleep, and set a timer $R_i$ for waking up in next round and at last go to sleep. If sensor node receives a mGOSLEEP message, it removes the neighbor sending that message out of its list L. All the sensors have to decide their status in the decision phase. After that, the active sensors perform the sensing task during the sensing phase. %The period the average @@ -239,11 +243,12 @@ check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e \begin{flushleft} \centering -\caption{Main characteristics of some coverage approaches in previous literatures.} +\caption{Main characteristics of some coverage approaches in literature.} \label{x11} \begin{tabular}{@{} cl*{13}c @{}} + & & \\ & & \multicolumn{10}{c}{Characteristics} \\[2ex] - \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ + \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds or Periods} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ \cmidrule[1pt]{2-14} @@ -331,11 +336,11 @@ check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e & \tiny X. Deng et al. (2005)~\cite{ref133} & \OK & & \OK & & \OK & & \OK & & \OK & & & &\\ -&\textbf{\textcolor{red}{ \tiny DiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & & &\textbf{\textcolor{red}{\OK}} & & \\ +&\textbf{\textcolor{red}{ \tiny DiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ -&\textbf{\textcolor{red}{ \tiny MuDiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} &\textbf{\textcolor{red}{\OK}} & & \\ +&\textbf{\textcolor{red}{ \tiny MuDiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} &\textbf{\textcolor{red}{\OK}} & & \\ -&\textbf{\textcolor{red}{ \tiny PeCO Protocol (2015)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & & &\textbf{\textcolor{red}{\OK}} & & \\ +&\textbf{\textcolor{red}{ \tiny PeCO Protocol (2015)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ \cmidrule[1pt]{2-14} \end{tabular} @@ -350,11 +355,11 @@ check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e \section{Conclusion} \label{ch2:sec:05} -This chapter describes some coverage proposed problems in the literature, with their assumptions and proposed solutions. -The coverage problem is considered as an essential requirement for many applications in WSNs because the better coverage of an area of interest provides better sensing measurements of the physical phenomenon. Therefore, many extensive researches have been focused on that problem. These researches aim at designing mechanisms that efficiently manage or schedule the sensors after deployment, since sensor nodes have a limited battery life. -Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead in WSNs. +This chapter describes some coverage problems in the literature, with their assumptions and proposed solutions. +The coverage is considered as an essential requirement for many applications in WSNs because the better the coverage of an area of interest, the better the sensing measurements of the physical phenomenon. Therefore, many extensive researches have been focused on that problem. These researches aim at designing mechanisms that efficiently manage or schedule the sensors after deployment, since sensor nodes have a limited battery life. +Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead. %Whatever the case, this would result in a lower lifetime coverage in WSNs. -As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. This hybrid approaches can provide a good quality coverage and prolong the network lifetime. +As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. A such hybrid approach can provide a good quality coverage and prolong the network lifetime. diff --git a/Thesis.toc b/Thesis.toc index 5062f88..4e3ad0a 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -41,11 +41,11 @@ \contentsline {section}{\numberline {1.11}Conclusion}{44}{section.1.11} \contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{45}{chapter.2} \contentsline {section}{\numberline {2.1}Introduction}{45}{section.2.1} -\contentsline {section}{\numberline {2.2}Centralized Algorithms}{47}{section.2.2} -\contentsline {section}{\numberline {2.3}Distributed Algorithms}{50}{section.2.3} -\contentsline {subsection}{\numberline {2.3.1}GAF}{52}{subsection.2.3.1} -\contentsline {subsection}{\numberline {2.3.2}DESK}{53}{subsection.2.3.2} -\contentsline {section}{\numberline {2.4}Conclusion}{56}{section.2.4} +\contentsline {section}{\numberline {2.2}Centralized Algorithms}{48}{section.2.2} +\contentsline {section}{\numberline {2.3}Distributed Algorithms}{51}{section.2.3} +\contentsline {subsection}{\numberline {2.3.1}Geographical Adaptive Fidelity (GAF)}{52}{subsection.2.3.1} +\contentsline {subsection}{\numberline {2.3.2}Distributed Energy-efficient Scheduling for K-coverage (DESK)}{54}{subsection.2.3.2} +\contentsline {section}{\numberline {2.4}Conclusion}{57}{section.2.4} \contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{59}{chapter.3} \contentsline {section}{\numberline {3.1}Introduction}{59}{section.3.1} \contentsline {section}{\numberline {3.2}Evaluation Tools}{59}{section.3.2} -- 2.39.5 From 133b4ccd131d5ec5facc732cc735a250d3bb81e3 Mon Sep 17 00:00:00 2001 From: ali Date: Sun, 26 Apr 2015 20:24:05 +0200 Subject: [PATCH 05/16] Update Today by Ali --- Resume.tex | 37 ++++++++++ Thesis.tex | 3 +- Thesis.toc | 209 +++++++++++++++++++++++++++-------------------------- 3 files changed, 144 insertions(+), 105 deletions(-) create mode 100644 Resume.tex diff --git a/Resume.tex b/Resume.tex new file mode 100644 index 0000000..15fa445 --- /dev/null +++ b/Resume.tex @@ -0,0 +1,37 @@ +\chapter*{Résumé \markboth{Résumé}{Résumé}} +\label{cha} +\addcontentsline{toc}{chapter}{Résumé} + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%% %% +%% Résumé %% +%% %% +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + + +\emph{ \begin{center} \Large Techniques d'Optimisation Couverture Distribuée pour Améliorer la Durée des Réseaux de Capteurs sans Fil \end{center}} +%\emph{ \begin{center} \large By \end{center}} +\emph{ \begin{center} \large Ali Kadhum Idrees \\ Université de Franche-Comt\'e, 2015 \end{center}} +%\emph{ \begin{center} \large The University of Franche-Comt\'e, 2015 \end{center}} +\emph{ \begin{center} \large Encadrants: Raphaël Couturier, Karine Deschinkel, and Michel Salomon \end{center}} + + +Réseaux de capteurs sans fil ont récemment reçu beaucoup d'attention de la recherche en raison de leur large gamme d'applications potentielles. Beaucoup de caractéristiques importantes sont fournis par les réseaux de capteurs qui les rendent différent des autres réseaux ad-hoc sans fil. Ces caractéristiques sont imposées beaucoup de limitations sur les réseaux de capteurs qui mèneraient à plusieurs défis dans le réseau. Ces défis pourraient inclure la couverture, contrôle de topologie, routage, la fusion de données, la sécurité, et bien d'autres. L'un des principaux défis de la recherche rencontrés dans les réseaux de capteurs sans fil est de préserver effectivement et en permanence la couverture d'une zone d'intérêt à surveiller, tout en empêchant simultanément autant que possible une défaillance du réseau en raison de nœuds de batterie appauvri. + +Dans cette thèse, nous nous concentrons fortement sur le problème de la zone de couverture, l'efficacité énergétique est également l'exigence avant tout. Nous avons examiné les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Les protocoles proposés distribués d'optimisation (y compris les algorithmes, les modèles, et la résolution des programmes entiers) doivent être protocoles économes en énergie. Adresser ce problème, cette thèse propose des approches en deux étapes. Tout d'abord, le champ de détection est divisée en plus petites sous-régions en utilisant le concept de la méthode de diviser pour régner. Deuxièmement, l'un de nos protocoles d'optimisation distribués proposées est distribuée et appliquée sur les nœuds de capteurs dans chaque sous-région afin d'optimiser la couverture et les performances de durée de vie. Dans cette thèse, trois protocoles d'optimisation de couverture sont proposés. Ces protocoles combinent deux techniques efficaces: élection du chef pour chaque sous-région, suivis par une planification fondée sur l'optimisation des décisions de planification d'activité du capteur pour chaque sous-région. + +Premièrement, nous proposons un protocole appelé Optimisation de couverture à vie (Distributed DILCO). Dans ce protocole, la durée de vie est divisée en périodes. Chaque période se compose de quatre phases: échange d'informations, leader électorales, de décision et de détection. Le processus de décision est +effectuée par un nœud leader, qui résout un programme entier afin de fournir un seul ensemble de nœuds de capteurs actifs de couverture pour assurer une couverture pendant la phase de détection de la période actuelle. + +Ensuite, nous abordons le problème d'une optimisation des passages répétés problème de la couverture de la zone dans les réseaux de capteurs. Le passages répétés Optimisation de couverture à vie (Distributed MuDiLCO) protocole est suggéré afin d'étudier la possibilité de fournir de multiples ensembles de couverture des capteurs pour la phase de détection. Protocole MuDiLCO travaille également en périodes pendant lesquelles ensembles de nœuds de capteurs sont programmés pour rester actif pour un certain nombre de tours pendant la phase de détection, pour assurer une couverture de manière à maximiser la durée de vie de réseaux de capteur sans fil. Le processus de décision est toujours effectuée par un nœud leader, qui résout un programme entier pour produire le meilleur représentant établit à être utilisé pendant les tours de la phase de détection. + + + +Enfin et surtout,, nous proposons une couverture Optimization (Peco) protocole basé périmètre qui est également réparti entre les nœuds de capteurs dans chaque nouveauté subregion.The de notre approche réside essentiellement dans la formulation d'un nouveau modèle d'optimisation mathématique basée sur le niveau de couverture de périmètre pour planifier les activités de capteurs. Un nouveau modèle de couverture du programme entier est résolu par le leader pendant la phase de décision de façon à fournir un seul ensemble de capteurs de couverture pour la phase de détection. + + +Simulations approfondies sont menées en utilisant la simulation à événements discrets OMNeT++ pour valider l'efficacité de chacun de nos protocoles proposés. Nous nous référons à la características capteur de méduse II de la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles sont fiables pour augmenter la couverture à vie de réseaux de capteur sans fil et améliorent les performances. + + +\textbf{MOTS-CLÉS:} Réseaux sans fil, les réseaux de capteurs, Zone de couverture, Durée de vie du réseau, optimisation, la planification, algorithmes distribués, Algorithmes centralisée, Robustesse, connectivité, l'efficacité énergétique, l'énergie réseau hétérogène, homogène réseau. + diff --git a/Thesis.tex b/Thesis.tex index e3e3839..94c4f1f 100644 --- a/Thesis.tex +++ b/Thesis.tex @@ -13,7 +13,8 @@ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -\include{Abstruct} +\include{Abstruct} +\include{Resume} %% Sommaire \tableofcontents \addcontentsline{toc}{chapter}{Table of Contents} diff --git a/Thesis.toc b/Thesis.toc index 4e3ad0a..2add095 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -1,106 +1,107 @@ \select@language {english} \contentsline {chapter}{Abstract}{1}{chapter*.1} -\contentsline {chapter}{Table of Contents}{6}{chapter*.2} -\contentsline {chapter}{List of Figures}{8}{chapter*.3} -\contentsline {chapter}{List of Tables}{9}{chapter*.4} -\contentsline {chapter}{List of Algorithms}{11}{chapter*.5} -\contentsline {chapter}{List of Abbreviations}{13}{chapter*.6} -\contentsline {chapter}{Introduction }{17}{chapter*.7} -\contentsline {section}{1. General Introduction }{17}{section*.8} -\contentsline {section}{2. Motivation of the Dissertation }{18}{section*.9} -\contentsline {section}{3. The Objective of this Dissertation}{18}{section*.10} -\contentsline {section}{4. Main Contributions of this Dissertation}{18}{section*.11} -\contentsline {section}{5. Dissertation Outline}{20}{section*.12} -\contentsline {part}{I\hspace {1em}Scientific Background}{21}{part.1} -\contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{23}{chapter.1} -\contentsline {section}{\numberline {1.1}Introduction}{23}{section.1.1} -\contentsline {section}{\numberline {1.2}Architecture}{24}{section.1.2} -\contentsline {section}{\numberline {1.3}Types of Wireless Sensor Networks}{26}{section.1.3} -\contentsline {section}{\numberline {1.4}Applications}{28}{section.1.4} -\contentsline {section}{\numberline {1.5}The Main Challenges}{31}{section.1.5} -\contentsline {section}{\numberline {1.6}Energy-Efficient Mechanisms of a working WSN}{33}{section.1.6} -\contentsline {subsection}{\numberline {1.6.1}Energy-Efficient Routing}{33}{subsection.1.6.1} -\contentsline {subsubsection}{\numberline {1.6.1.1}Routing Metric based on Residual Energy}{33}{subsubsection.1.6.1.1} -\contentsline {subsubsection}{\numberline {1.6.1.2}Multipath Routing}{33}{subsubsection.1.6.1.2} -\contentsline {subsection}{\numberline {1.6.2}Cluster Architecture}{34}{subsection.1.6.2} -\contentsline {subsection}{\numberline {1.6.3}Scheduling Schemes}{34}{subsection.1.6.3} -\contentsline {subsubsection}{\numberline {1.6.3.1}Wake up Scheduling Schemes}{34}{subsubsection.1.6.3.1} -\contentsline {subsubsection}{\numberline {1.6.3.2}Topology Control Schemes}{37}{subsubsection.1.6.3.2} -\contentsline {subsection}{\numberline {1.6.4}Data-Driven Schemes}{37}{subsection.1.6.4} -\contentsline {subsubsection}{\numberline {1.6.4.1}Data Reduction Schemes}{38}{subsubsection.1.6.4.1} -\contentsline {subsubsection}{\numberline {1.6.4.2}Energy Efficient Data Acquisition Schemes}{38}{subsubsection.1.6.4.2} -\contentsline {subsection}{\numberline {1.6.5}Battery Repletion}{38}{subsection.1.6.5} -\contentsline {subsection}{\numberline {1.6.6}Radio Optimization}{38}{subsection.1.6.6} -\contentsline {subsection}{\numberline {1.6.7}Relay nodes and Sink Mobility}{39}{subsection.1.6.7} -\contentsline {subsubsection}{\numberline {1.6.7.1}Relay node placement}{39}{subsubsection.1.6.7.1} -\contentsline {subsubsection}{\numberline {1.6.7.2}Sink Mobility}{39}{subsubsection.1.6.7.2} -\contentsline {section}{\numberline {1.7}Network Lifetime}{39}{section.1.7} -\contentsline {section}{\numberline {1.8}Coverage in Wireless Sensor Networks }{40}{section.1.8} -\contentsline {section}{\numberline {1.9}Design Issues for Coverage Problems}{42}{section.1.9} -\contentsline {section}{\numberline {1.10}Energy Consumption Model}{43}{section.1.10} -\contentsline {section}{\numberline {1.11}Conclusion}{44}{section.1.11} -\contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{45}{chapter.2} -\contentsline {section}{\numberline {2.1}Introduction}{45}{section.2.1} -\contentsline {section}{\numberline {2.2}Centralized Algorithms}{48}{section.2.2} -\contentsline {section}{\numberline {2.3}Distributed Algorithms}{51}{section.2.3} -\contentsline {subsection}{\numberline {2.3.1}Geographical Adaptive Fidelity (GAF)}{52}{subsection.2.3.1} -\contentsline {subsection}{\numberline {2.3.2}Distributed Energy-efficient Scheduling for K-coverage (DESK)}{54}{subsection.2.3.2} -\contentsline {section}{\numberline {2.4}Conclusion}{57}{section.2.4} -\contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{59}{chapter.3} -\contentsline {section}{\numberline {3.1}Introduction}{59}{section.3.1} -\contentsline {section}{\numberline {3.2}Evaluation Tools}{59}{section.3.2} -\contentsline {subsection}{\numberline {3.2.1}Testbed Tools}{60}{subsection.3.2.1} -\contentsline {subsection}{\numberline {3.2.2}Simulation Tools}{61}{subsection.3.2.2} -\contentsline {section}{\numberline {3.3}Optimization Solvers}{66}{section.3.3} -\contentsline {section}{\numberline {3.4}Conclusion}{69}{section.3.4} -\contentsline {part}{II\hspace {1em}Contributions}{71}{part.2} -\contentsline {chapter}{\numberline {4}Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{73}{chapter.4} -\contentsline {section}{\numberline {4.1}Introduction}{73}{section.4.1} -\contentsline {section}{\numberline {4.2}Description of the DiLCO Protocol}{74}{section.4.2} -\contentsline {subsection}{\numberline {4.2.1}Assumptions and Network Model}{74}{subsection.4.2.1} -\contentsline {subsection}{\numberline {4.2.2}Primary Point Coverage Model}{75}{subsection.4.2.2} -\contentsline {subsection}{\numberline {4.2.3}Main Idea}{76}{subsection.4.2.3} -\contentsline {subsubsection}{\numberline {4.2.3.1}Information Exchange Phase}{77}{subsubsection.4.2.3.1} -\contentsline {subsubsection}{\numberline {4.2.3.2}Leader Election Phase}{77}{subsubsection.4.2.3.2} -\contentsline {subsubsection}{\numberline {4.2.3.3}Decision phase}{77}{subsubsection.4.2.3.3} -\contentsline {subsubsection}{\numberline {4.2.3.4}Sensing phase}{77}{subsubsection.4.2.3.4} -\contentsline {section}{\numberline {4.3}Primary Points based Coverage Problem Formulation}{78}{section.4.3} -\contentsline {section}{\numberline {4.4}Simulation Results and Analysis}{80}{section.4.4} -\contentsline {subsection}{\numberline {4.4.1}Simulation Framework}{80}{subsection.4.4.1} -\contentsline {subsection}{\numberline {4.4.2}Modeling Language and Optimization Solver}{80}{subsection.4.4.2} -\contentsline {subsection}{\numberline {4.4.3}Energy Consumption Model}{80}{subsection.4.4.3} -\contentsline {subsection}{\numberline {4.4.4}Performance Metrics}{81}{subsection.4.4.4} -\contentsline {subsection}{\numberline {4.4.5}Performance Analysis for Different Subregions}{82}{subsection.4.4.5} -\contentsline {subsection}{\numberline {4.4.6}Performance Analysis for Primary Point Models}{88}{subsection.4.4.6} -\contentsline {subsection}{\numberline {4.4.7}Performance Comparison with other Approaches}{93}{subsection.4.4.7} -\contentsline {section}{\numberline {4.5}Conclusion}{99}{section.4.5} -\contentsline {chapter}{\numberline {5}Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{101}{chapter.5} -\contentsline {section}{\numberline {5.1}Introduction}{101}{section.5.1} -\contentsline {section}{\numberline {5.2}MuDiLCO Protocol Description}{102}{section.5.2} -\contentsline {subsection}{\numberline {5.2.1}Background Idea and Algorithm}{102}{subsection.5.2.1} -\contentsline {section}{\numberline {5.3}Primary Points based Multiround Coverage Problem Formulation}{103}{section.5.3} -\contentsline {section}{\numberline {5.4}Experimental Study and Analysis}{105}{section.5.4} -\contentsline {subsection}{\numberline {5.4.1}Simulation Setup}{105}{subsection.5.4.1} -\contentsline {subsection}{\numberline {5.4.2}Metrics}{106}{subsection.5.4.2} -\contentsline {subsection}{\numberline {5.4.3}Results Analysis and Comparison }{107}{subsection.5.4.3} -\contentsline {section}{\numberline {5.5}Conclusion}{112}{section.5.5} -\contentsline {chapter}{\numberline {6}Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks}{115}{chapter.6} -\contentsline {section}{\numberline {6.1}Introduction}{115}{section.6.1} -\contentsline {section}{\numberline {6.2}The PeCO Protocol Description}{116}{section.6.2} -\contentsline {subsection}{\numberline {6.2.1}Assumptions and Models}{116}{subsection.6.2.1} -\contentsline {subsection}{\numberline {6.2.2}The Main Idea}{119}{subsection.6.2.2} -\contentsline {subsection}{\numberline {6.2.3}PeCO Protocol Algorithm}{119}{subsection.6.2.3} -\contentsline {section}{\numberline {6.3}Perimeter-based Coverage Problem Formulation}{120}{section.6.3} -\contentsline {section}{\numberline {6.4}Performance Evaluation and Analysis}{122}{section.6.4} -\contentsline {subsection}{\numberline {6.4.1}Simulation Settings}{122}{subsection.6.4.1} -\contentsline {subsection}{\numberline {6.4.2}Simulation Results}{123}{subsection.6.4.2} -\contentsline {subsubsection}{\numberline {6.4.2.1}Coverage Ratio}{124}{subsubsection.6.4.2.1} -\contentsline {subsubsection}{\numberline {6.4.2.2}Active Sensors Ratio}{124}{subsubsection.6.4.2.2} -\contentsline {subsubsection}{\numberline {6.4.2.3}The Energy Consumption}{125}{subsubsection.6.4.2.3} -\contentsline {subsubsection}{\numberline {6.4.2.4}The Network Lifetime}{125}{subsubsection.6.4.2.4} -\contentsline {section}{\numberline {6.5}Conclusion}{128}{section.6.5} -\contentsline {part}{III\hspace {1em}Conclusion and Perspectives}{129}{part.3} -\contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{131}{chapter.7} -\contentsline {section}{\numberline {7.1}Conclusion}{131}{section.7.1} -\contentsline {section}{\numberline {7.2}Perspectives}{132}{section.7.2} -\contentsline {part}{Bibliographie}{148}{chapter*.13} +\contentsline {chapter}{R\IeC {\'e}sum\IeC {\'e}}{3}{chapter*.2} +\contentsline {chapter}{Table of Contents}{8}{chapter*.3} +\contentsline {chapter}{List of Figures}{10}{chapter*.4} +\contentsline {chapter}{List of Tables}{11}{chapter*.5} +\contentsline {chapter}{List of Algorithms}{13}{chapter*.6} +\contentsline {chapter}{List of Abbreviations}{15}{chapter*.7} +\contentsline {chapter}{Introduction }{19}{chapter*.8} +\contentsline {section}{1. General Introduction }{19}{section*.9} +\contentsline {section}{2. Motivation of the Dissertation }{20}{section*.10} +\contentsline {section}{3. The Objective of this Dissertation}{20}{section*.11} +\contentsline {section}{4. Main Contributions of this Dissertation}{20}{section*.12} +\contentsline {section}{5. Dissertation Outline}{22}{section*.13} +\contentsline {part}{I\hspace {1em}Scientific Background}{23}{part.1} +\contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{25}{chapter.1} +\contentsline {section}{\numberline {1.1}Introduction}{25}{section.1.1} +\contentsline {section}{\numberline {1.2}Architecture}{26}{section.1.2} +\contentsline {section}{\numberline {1.3}Types of Wireless Sensor Networks}{28}{section.1.3} +\contentsline {section}{\numberline {1.4}Applications}{30}{section.1.4} +\contentsline {section}{\numberline {1.5}The Main Challenges}{33}{section.1.5} +\contentsline {section}{\numberline {1.6}Energy-Efficient Mechanisms of a working WSN}{35}{section.1.6} +\contentsline {subsection}{\numberline {1.6.1}Energy-Efficient Routing}{35}{subsection.1.6.1} +\contentsline {subsubsection}{\numberline {1.6.1.1}Routing Metric based on Residual Energy}{35}{subsubsection.1.6.1.1} +\contentsline {subsubsection}{\numberline {1.6.1.2}Multipath Routing}{35}{subsubsection.1.6.1.2} +\contentsline {subsection}{\numberline {1.6.2}Cluster Architecture}{36}{subsection.1.6.2} +\contentsline {subsection}{\numberline {1.6.3}Scheduling Schemes}{36}{subsection.1.6.3} +\contentsline {subsubsection}{\numberline {1.6.3.1}Wake up Scheduling Schemes}{36}{subsubsection.1.6.3.1} +\contentsline {subsubsection}{\numberline {1.6.3.2}Topology Control Schemes}{39}{subsubsection.1.6.3.2} +\contentsline {subsection}{\numberline {1.6.4}Data-Driven Schemes}{39}{subsection.1.6.4} +\contentsline {subsubsection}{\numberline {1.6.4.1}Data Reduction Schemes}{40}{subsubsection.1.6.4.1} +\contentsline {subsubsection}{\numberline {1.6.4.2}Energy Efficient Data Acquisition Schemes}{40}{subsubsection.1.6.4.2} +\contentsline {subsection}{\numberline {1.6.5}Battery Repletion}{40}{subsection.1.6.5} +\contentsline {subsection}{\numberline {1.6.6}Radio Optimization}{40}{subsection.1.6.6} +\contentsline {subsection}{\numberline {1.6.7}Relay nodes and Sink Mobility}{41}{subsection.1.6.7} +\contentsline {subsubsection}{\numberline {1.6.7.1}Relay node placement}{41}{subsubsection.1.6.7.1} +\contentsline {subsubsection}{\numberline {1.6.7.2}Sink Mobility}{41}{subsubsection.1.6.7.2} +\contentsline {section}{\numberline {1.7}Network Lifetime}{41}{section.1.7} +\contentsline {section}{\numberline {1.8}Coverage in Wireless Sensor Networks }{42}{section.1.8} +\contentsline {section}{\numberline {1.9}Design Issues for Coverage Problems}{44}{section.1.9} +\contentsline {section}{\numberline {1.10}Energy Consumption Model}{45}{section.1.10} +\contentsline {section}{\numberline {1.11}Conclusion}{46}{section.1.11} +\contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{47}{chapter.2} +\contentsline {section}{\numberline {2.1}Introduction}{47}{section.2.1} +\contentsline {section}{\numberline {2.2}Centralized Algorithms}{50}{section.2.2} +\contentsline {section}{\numberline {2.3}Distributed Algorithms}{53}{section.2.3} +\contentsline {subsection}{\numberline {2.3.1}Geographical Adaptive Fidelity (GAF)}{54}{subsection.2.3.1} +\contentsline {subsection}{\numberline {2.3.2}Distributed Energy-efficient Scheduling for K-coverage (DESK)}{56}{subsection.2.3.2} +\contentsline {section}{\numberline {2.4}Conclusion}{59}{section.2.4} +\contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{61}{chapter.3} +\contentsline {section}{\numberline {3.1}Introduction}{61}{section.3.1} +\contentsline {section}{\numberline {3.2}Evaluation Tools}{61}{section.3.2} +\contentsline {subsection}{\numberline {3.2.1}Testbed Tools}{62}{subsection.3.2.1} +\contentsline {subsection}{\numberline {3.2.2}Simulation Tools}{63}{subsection.3.2.2} +\contentsline {section}{\numberline {3.3}Optimization Solvers}{68}{section.3.3} +\contentsline {section}{\numberline {3.4}Conclusion}{71}{section.3.4} +\contentsline {part}{II\hspace {1em}Contributions}{73}{part.2} +\contentsline {chapter}{\numberline {4}Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{75}{chapter.4} +\contentsline {section}{\numberline {4.1}Introduction}{75}{section.4.1} +\contentsline {section}{\numberline {4.2}Description of the DiLCO Protocol}{76}{section.4.2} +\contentsline {subsection}{\numberline {4.2.1}Assumptions and Network Model}{76}{subsection.4.2.1} +\contentsline {subsection}{\numberline {4.2.2}Primary Point Coverage Model}{77}{subsection.4.2.2} +\contentsline {subsection}{\numberline {4.2.3}Main Idea}{78}{subsection.4.2.3} +\contentsline {subsubsection}{\numberline {4.2.3.1}Information Exchange Phase}{79}{subsubsection.4.2.3.1} +\contentsline {subsubsection}{\numberline {4.2.3.2}Leader Election Phase}{79}{subsubsection.4.2.3.2} +\contentsline {subsubsection}{\numberline {4.2.3.3}Decision phase}{79}{subsubsection.4.2.3.3} +\contentsline {subsubsection}{\numberline {4.2.3.4}Sensing phase}{79}{subsubsection.4.2.3.4} +\contentsline {section}{\numberline {4.3}Primary Points based Coverage Problem Formulation}{80}{section.4.3} +\contentsline {section}{\numberline {4.4}Simulation Results and Analysis}{82}{section.4.4} +\contentsline {subsection}{\numberline {4.4.1}Simulation Framework}{82}{subsection.4.4.1} +\contentsline {subsection}{\numberline {4.4.2}Modeling Language and Optimization Solver}{82}{subsection.4.4.2} +\contentsline {subsection}{\numberline {4.4.3}Energy Consumption Model}{82}{subsection.4.4.3} +\contentsline {subsection}{\numberline {4.4.4}Performance Metrics}{83}{subsection.4.4.4} +\contentsline {subsection}{\numberline {4.4.5}Performance Analysis for Different Subregions}{84}{subsection.4.4.5} +\contentsline {subsection}{\numberline {4.4.6}Performance Analysis for Primary Point Models}{90}{subsection.4.4.6} +\contentsline {subsection}{\numberline {4.4.7}Performance Comparison with other Approaches}{95}{subsection.4.4.7} +\contentsline {section}{\numberline {4.5}Conclusion}{101}{section.4.5} +\contentsline {chapter}{\numberline {5}Multiround Distributed Lifetime Coverage Optimization Protocol in Wireless Sensor Networks}{103}{chapter.5} +\contentsline {section}{\numberline {5.1}Introduction}{103}{section.5.1} +\contentsline {section}{\numberline {5.2}MuDiLCO Protocol Description}{104}{section.5.2} +\contentsline {subsection}{\numberline {5.2.1}Background Idea and Algorithm}{104}{subsection.5.2.1} +\contentsline {section}{\numberline {5.3}Primary Points based Multiround Coverage Problem Formulation}{105}{section.5.3} +\contentsline {section}{\numberline {5.4}Experimental Study and Analysis}{107}{section.5.4} +\contentsline {subsection}{\numberline {5.4.1}Simulation Setup}{107}{subsection.5.4.1} +\contentsline {subsection}{\numberline {5.4.2}Metrics}{108}{subsection.5.4.2} +\contentsline {subsection}{\numberline {5.4.3}Results Analysis and Comparison }{109}{subsection.5.4.3} +\contentsline {section}{\numberline {5.5}Conclusion}{114}{section.5.5} +\contentsline {chapter}{\numberline {6}Perimeter-based Coverage Optimization to Improve Lifetime in Wireless Sensor Networks}{117}{chapter.6} +\contentsline {section}{\numberline {6.1}Introduction}{117}{section.6.1} +\contentsline {section}{\numberline {6.2}The PeCO Protocol Description}{118}{section.6.2} +\contentsline {subsection}{\numberline {6.2.1}Assumptions and Models}{118}{subsection.6.2.1} +\contentsline {subsection}{\numberline {6.2.2}The Main Idea}{121}{subsection.6.2.2} +\contentsline {subsection}{\numberline {6.2.3}PeCO Protocol Algorithm}{121}{subsection.6.2.3} +\contentsline {section}{\numberline {6.3}Perimeter-based Coverage Problem Formulation}{122}{section.6.3} +\contentsline {section}{\numberline {6.4}Performance Evaluation and Analysis}{124}{section.6.4} +\contentsline {subsection}{\numberline {6.4.1}Simulation Settings}{124}{subsection.6.4.1} +\contentsline {subsection}{\numberline {6.4.2}Simulation Results}{125}{subsection.6.4.2} +\contentsline {subsubsection}{\numberline {6.4.2.1}Coverage Ratio}{126}{subsubsection.6.4.2.1} +\contentsline {subsubsection}{\numberline {6.4.2.2}Active Sensors Ratio}{126}{subsubsection.6.4.2.2} +\contentsline {subsubsection}{\numberline {6.4.2.3}The Energy Consumption}{127}{subsubsection.6.4.2.3} +\contentsline {subsubsection}{\numberline {6.4.2.4}The Network Lifetime}{127}{subsubsection.6.4.2.4} +\contentsline {section}{\numberline {6.5}Conclusion}{130}{section.6.5} +\contentsline {part}{III\hspace {1em}Conclusion and Perspectives}{131}{part.3} +\contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{133}{chapter.7} +\contentsline {section}{\numberline {7.1}Conclusion}{133}{section.7.1} +\contentsline {section}{\numberline {7.2}Perspectives}{134}{section.7.2} +\contentsline {part}{Bibliographie}{150}{chapter*.14} -- 2.39.5 From bb06163c8d122bfd6baf424927a264670a40b29e Mon Sep 17 00:00:00 2001 From: ali Date: Fri, 1 May 2015 01:29:07 +0200 Subject: [PATCH 06/16] Update by Ali --- CHAPITRE_01.tex | 56 ++++++++++++++++++++++++------------------------- 1 file changed, 28 insertions(+), 28 deletions(-) diff --git a/CHAPITRE_01.tex b/CHAPITRE_01.tex index ec27e65..5ba6c61 100644 --- a/CHAPITRE_01.tex +++ b/CHAPITRE_01.tex @@ -13,7 +13,7 @@ \label{ch1:sec:01} %The wireless networking has received more attention and fast growth in the last decade. In the last decade, wireless networking has became a major component of the global network infrastructure. -More precisely, the growing demand for the use of wireless applications and the continuous arrival of wireless devices such as portable computers, cellular phones, and Personal Digital Assistants (PDAs) have led to develop different infrastructures of wireless networks. The wireless networks can be classified into two classes based on the network architecture~\cite{ref154,ref155}: Infrastructure-based networks that consist of a fixed network structure such as cellular networks and Wireless Local-Area Networks +More precisely, the growing demand for the use of wireless applications and the continuous arrival of wireless devices such as portable computers, cellular phones, and Personal Digital Assistants (PDAs) have led to develop different infrastructures of wireless networks. Wireless networks can be classified into two classes based on the network architecture~\cite{ref154,ref155}: Infrastructure-based networks that consist of a fixed network structure such as cellular networks and Wireless Local-Area Networks (WLANs); and Infrastructureless networks that are constructed dynamically by the cooperation of the wireless nodes in the network, where each node is capable of sending packets and taking decisions based on the network status. Examples of such type of networks include mobile ad hoc networks and wireless sensor networks. Figure~\ref{WNT} shows the taxonomy of wireless networks. \begin{figure}[h!] @@ -28,12 +28,12 @@ More precisely, the growing demand for the use of wireless applications and the %WSNs are considered as one of the most researched fields in the last decade due to the extensive research in this discipline. Wireless Sensor Networks (WSNs) represent a special case of the Ad Hoc networks, resulting from recent advances in wireless networking, Micro-Electro-Mechanical Systems (MEMS), and embedded computing technologies, which have led to construct low-cost, small-sized, and low-power sensor nodes. %These sensor nodes can perform detection, computation, and data communication of surrounding environment. -A WSN includes a large number of sensor nodes that can sense, process, and transmit data over a wireless communication. The sensor nodes communicate with each other by using multi-hop wireless communications and cooperate together to monitor the area of interest. Measured data are reported to a monitoring center called base station or sink for further analysis~\cite{ref1,ref2}. The WSN receives the orders from the end user by means of the sink. These orders specify data aggregation, computation, and delivery missions to wireless sensor nodes, after that the sensed measurements are received from the WSN by the sink~\cite{ref3}. The cooperation among wireless sensor nodes in WSNs has several advantages over traditional wireless ad-hoc networks, like self-organization, rapid deployment, flexibility, and inherent intelligent-processing capability~\cite{ref5}. +A WSN includes a large number of sensor nodes that can sense, process, and transmit data over a wireless communication. The sensor nodes communicate with each other by using multi-hop wireless communications and cooperate together to monitor the area of interest. Measured data are reported to a monitoring center called base station or sink for further analysis~\cite{ref1,ref2}. A WSN receives the orders from the end user by means of the sink. These orders specify data aggregation, computation, and delivery missions to wireless sensor nodes, after that the sensed measurements are received from the WSN by the sink~\cite{ref3}. The cooperation among wireless sensor nodes in WSNs has several advantages over traditional wireless ad-hoc networks, like self-organization, rapid deployment, flexibility, and inherent intelligent-processing capability~\cite{ref5}. \section{Architecture} \label{ch1:sec:02} -A typical WSN architecture consists of in a set of a huge number of wireless sensor nodes, which are capable of sensing the surrounded physical phenomenon such as fire in the forest (see~figure~\ref{wsn}), and then send the sensed data to a sink node. One or more sink in WSN are responsible for collecting and processing the received sensed data, and making them available through the Internet to the end-user. +A typical WSN architecture consists in a set of a huge number of wireless sensor nodes, which are capable of sensing the surrounded physical phenomenon such as fire in the forest (see~figure~\ref{wsn}), and then send the sensed data to a sink node. One or more sinks in WSN are responsible for collecting and processing the received sensed data, and making them available through the Internet to the end-user. The basic element is a wireless sensor node that is composed of four major units~\cite{ref17,ref18}: sensing, computation, communication, and power. %In addition, there are three optional units, which can be combined with the sensor node such as localization system, mobilizer, and power generator. @@ -75,12 +75,12 @@ Furthermore, additional components can be incorporated into wireless sensor node \label{wsn} \end{figure} -The sensor node use a software layer called, Operating System (OS), is logically locates between the node's hardware and the application layer~\cite{ref18}. The OS enables the applications to interact with hardware resources, to schedule and prioritize tasks, memory management, power management, file management, networking, and to arbitrate between contending applications and services that attempt to reserve resources. The TinyOS has been used as an operating system in wireless sensor node. It is developed by the university of California, Berkeley and designed to work on platforms with limited storage and processing power. +Sensor nodes use a software layer called, Operating System (OS), which is logically between the node's hardware and the application layer~\cite{ref18}. The OS enables the applications to interact with hardware resources, to schedule and prioritize tasks, memory management, power management, file management, networking, and to arbitrate between contending applications and services that attempt to reserve resources. The TinyOS has been used as an operating system in wireless sensor node. It is developed by the university of California, Berkeley and designed to work on platforms with limited storage and processing power. \section{Types of Wireless Sensor Networks} \label{ch1:sec:03} -According to the physical phenomena for which the WSN is developed, WSNs can be deployed on the ground, underground, or underwater. WSNs suffer from different conditions and challenges. WSNs can be classified into six types, thus among which five types are presented in~\cite{ref4,ref5}. Figure~\ref{wsnt} gives examples of different WSNs types. +According to the physical phenomena for which the WSN is developed, WSNs can be deployed on the ground, underground, or underwater. WSNs suffer from different conditions and challenges. WSNs can be classified into six types, among which five types are presented in~\cite{ref4,ref5}. Figure~\ref{wsnt} gives examples of different WSNs types. \begin{figure}[h!] \centering \includegraphics[scale=0.5]{Figures/ch1/typesWSN.pdf} @@ -91,7 +91,7 @@ According to the physical phenomena for which the WSN is developed, WSNs can be \begin{enumerate}[(I)] \item \textbf{Terrestrial WSNs:} -The wireless sensor nodes are deployed over the land constructing a network of hundred to thousand of sensor devices. Several applications use terrestrial WSNs for physical environmental sensing and monitoring, industrial monitoring, and surface explorations. The main challenges in this type of WSNs are to ensure coverage and connectivity with removing redundancy, energy-efficient routing, data communication reduction, balancing energy consumption, energy-efficient data aggregation. +Wireless sensor nodes are deployed over the land constructing a network of hundred to thousand of sensor devices. Several applications use terrestrial WSNs for physical environmental sensing and monitoring, industrial monitoring, and surface explorations. The main challenges in this type of WSNs are to ensure coverage and connectivity with removing redundancy, energy-efficient routing, data communication reduction, balancing energy consumption, energy-efficient data aggregation. %This dissertation focuses on this type of WSNs. \item \textbf{Underground WSNs:} @@ -106,7 +106,7 @@ They consist of inexpensive wireless sensor nodes supplied with CMOS (Complement \item \textbf{Mobile WSNs:} Such a network is composed of mobile sensor nodes that can move autonomously so as to self-organize~\cite{ref16}. %self-moving and reacting for the physical phenomena~\cite{ref16}. The mobile sensor node is self-organized and it is capable of replacing its position autonomously. In addition, it is able to sense, process, and communicate with other mobile sensors. -Many challenges should be faced in mobile WSNs such as maintaining a sufficient sensing coverage and connectivity; the self-organization; the navigation and control of mobile sensors; mobility management; processing in WSN; location determination with mobility; and minimizing the energy consumption especially during the movement. The mobile WSN applications are environment, habitat, and underwater monitoring; target tracking; military surveillance; search and rescue. The mobility in WSNs improves the network coverage of the monitored area especially after the initial random deployment becuase they can relocate themselves to fill coverage holes \cite{ref234,ref235}. +Many challenges should be faced in mobile WSNs such as maintaining a sufficient sensing coverage and connectivity; the self-organization; the navigation and control of mobile sensors; mobility management; processing in WSN; location determination with mobility; and minimizing the energy consumption especially during the movement. The mobile WSN applications are environment, habitat, and underwater monitoring; target tracking; military surveillance; search and rescue. The mobility in WSNs improves the network coverage of the monitored area especially after the initial random deployment because they can relocate themselves to fill coverage holes \cite{ref234,ref235}. \item \textbf{Flying WSNs:} This kind of WSN consists of low-cost wireless sensor nodes, which are embedded in Micro Aerial Vehicles (MAVs). They can fly autonomously or can be operated remotely without intervention of any human personnel~\cite{ref6,ref7}. The general objective of this type of WSN is to retrieve information from some inaccessible locations. For example, establishing an ad hoc network connection between rescuers and disaster victims over airborne relays or surveying an area from the air. A flying WSN provides a remote sensing and wireless networking platform that collect the data from local sensors or other sources, and send the collected information over airborne wireless relays to a ground station. Using Flying WSNs have led to new developments for both military and civilian applications due to their flexibility, versatility, easy installation, and the operating low-cost \cite{ref8}. The applications are search and destroy operations, disaster monitoring, relay for ad hoc networks, wind estimation, managing wildfire, border surveillance, remote sensing, and traffic monitoring. The main challenges are constructing a lightweight MAV which is capable of flight; the wireless communication; designing software protocols to achieve semi-autonomous flight; and combining all the subsystems like propulsion, flight control, and wireless networking into a flying WSN. @@ -127,14 +127,14 @@ In this section, we describe different academic and commercial applications. A W \begin{enumerate}[(I)] -\item \textbf{Health-care Applications:} There is an increasing interest and extensive research in this domain. Two types of health-care systems are recognized~\cite{ref22}: vital status monitoring and remote health-care surveillance. In vital status monitoring applications, sick persons are wearing the sensors in order to oversee their health state and to allow medical staff to monitor and control the patient's status expeditiously. The most general used vital signs are ECG, pulse oximetry, body temperature, heart rate, and blood pressure~\cite{ref27}. These applications include mass-casualty disaster monitoring, vital sign monitoring in hospitals, and sudden fall or epilepsy seizure detection. On the other hand, remote health-care surveillance refers to health services that do not require continuous existence of health care. These applications include elderly monitoring, providing support to a physically impaired person, gather clinically relevant information for rehabilitation supervision~\cite{ref28}, location tracking, and medication intake monitoring~\cite{ref27}. +\item \textbf{Health-care Applications:} There is an increasing interest and extensive research in this domain. Two types of health-care systems are recognized~\cite{ref22}: vital status monitoring and remote health-care surveillance. In vital status monitoring applications, sick persons are wearing the sensors in order to oversee their health state and to allow medical staff to monitor and control the patient's status expeditiously. The most general used vital signs are electrocardiogram (ECG), pulse oximetry, body temperature, heart rate, and blood pressure~\cite{ref27}. These applications include mass-casualty disaster monitoring, vital sign monitoring in hospitals, and sudden fall or epilepsy seizure detection. On the other hand, remote health-care surveillance refers to health services that do not require continuous existence of health care. These applications include elderly monitoring, providing support to a physically impaired person, gather clinically relevant information for rehabilitation supervision~\cite{ref28}, location tracking, and medication intake monitoring~\cite{ref27}. \item \textbf{ Environment and agriculture Applications} \indent Several WSNs applications have been developed for precision agriculture, cattle monitoring, and environmental monitoring. \indent Precision agriculture refers to the science of using innovative and modern technologies to improve the crop production. WSNs are the main technology for developing precision agriculture~\cite{ref29}. This technology contributes to increasing the agricultural yields, improving quality, and reducing costs whilst decreasing the damaging impact on the environment. The wireless sensors are distributed over the target field so as to monitor the main parameters such as soil moisture, atmospheric temperature, and create a decision support system \cite{ref22}. -The wireless sensors can be used in agricultural services like Irrigation, fertilization, pest control, animal and pastures monitoring, horticulture (e.g., greenhouse and viticulture)~\cite{ref30}. For instance, in cattle monitoring applications, the WSN is used to livestock control and monitoring such as virtual fencing for extensive grazing systems, animal behavior study, health monitoring, to detect disease breakouts, to localize them, and to control end-product quality (meat, milk). +The wireless sensors can be used in agricultural services like irrigation, fertilization, pest control, animal and pastures monitoring, horticulture (e.g., greenhouse and viticulture)~\cite{ref30}. For instance, in cattle monitoring applications, the WSN is used to livestock control and monitoring such as virtual fencing for extensive grazing systems, animal behavior study, health monitoring, to detect disease breakouts, to localize them, and to control end-product quality (meat, milk). \indent Various WSN applications for environmental monitoring have been used in coastline erosion, air quality monitoring, safe drinking water, and contamination control~\cite{ref30,ref22}. @@ -156,7 +156,7 @@ The fast development in the domain of Intelligent Transport Systems (ITS) rangin \item \textbf{Industry Applications: Manufacturing and Smart Grids:} -The most significant goal for many companies is the automation of controlling and monitoring systems in many applications such as manufacturing, water treatment, electrical power distribution, and oil and gas refining. In that case WSNs are incorporated in Supervisory Control and Data Acquisition (SCADA) systems and smart grids~\cite{ref22}.A SCADA system is a computer software by which industrial processes in factories are controlled and supervised. The wireless sensors are used with actuators to control the factory, to detect of liquid/gas leakages, and for inventory management. These applications are needed for precise monitoring of temperature, shock, and noise factors in remote locations such as tanks, turbine engines, or pipelines. In Smart Grids, the goal is to supervise the power supply and depletion operation. The main applications in smart grid include: sensing the relevant parameters affecting power output (pressure, humidity, wind orientation, radiation, etc.); control of turbines, motors and underground cables; home energy management; and remote detection of faulty components. +The most significant goal for many companies is the automation of controlling and monitoring systems in many applications such as manufacturing, water treatment, electrical power distribution, and oil and gas refining. In that case WSNs are incorporated in Supervisory Control and Data Acquisition (SCADA) systems and smart grids~\cite{ref22}. A SCADA system is a computer software by which industrial processes in factories are controlled and supervised. The wireless sensors are used with actuators to control the factory, to detect of liquid/gas leakages, and for inventory management. These applications are needed for precise monitoring of temperature, shock, and noise factors in remote locations such as tanks, turbine engines, or pipelines. In Smart Grids, the goal is to supervise the power supply and depletion operation. The main applications in smart grid include: sensing the relevant parameters affecting power output (pressure, humidity, wind orientation, radiation, etc.); control of turbines, motors and underground cables; home energy management; and remote detection of faulty components. \end{enumerate} %\section{Protocol Design Requirements} @@ -176,9 +176,9 @@ The most significant goal for many companies is the automation of controlling an \item \textbf{Routing:} It represents one of the important problems in WSNs that needs to be solved efficiently. The limited resources of WSNs and the impacts of wireless communication lead to a big challenge in ensuring energy-efficient routing. However, it is not enough to use the shortest path to route the packets among the sensor nodes toward the sink. It is necessary to design an energy-efficient routing protocol that considers the remaining energy of the sensor node during taking the decision to route the packet to the next hop toward the destination. This participates in energy conservation and balancing among the sensor nodes in WSNs. \item \textbf{Autonomous and Distributed Management:} -Since the nature of many WSN applications induce a deployment in a remote or hostile environment, it is important that the wireless sensor nodes work in an autonomous and distributed way to communicate and cooperate, without any human intervention since the maintenance or the repair may be difficult. %The distributed management consumes less energy because it is based on only local information from the neighboring sensor nodes; moreover, it does not give the optimal solution. Therefore, the main challenge is how to apply a distributed management in WSNs and in the same time ensuring an optimal or near optimal solution. +Since the nature of many WSN applications induces a deployment in a remote or hostile environment, it is important that the wireless sensor nodes work in an autonomous and distributed way to communicate and cooperate, without any human intervention since the maintenance or the repair may be difficult. %The distributed management consumes less energy because it is based on only local information from the neighboring sensor nodes; moreover, it does not give the optimal solution. Therefore, the main challenge is how to apply a distributed management in WSNs and in the same time ensuring an optimal or near optimal solution. -\item \textbf{Scalability:} Many physical phenomenons require the deployment of a dense WSN. A large number of sensor nodes maybe needed for different reasons such as the huge size of the sensed area, the reliability requirement, or network lifetime prolongation. It is necessary that the proposed protocols for WSNs are scalable for these large number of sensor nodes in order to achieve their tasks efficiently. +\item \textbf{Scalability:} Many physical phenomenons require the deployment of a dense WSN. A large number of sensor nodes may be needed for different reasons such as the huge size of the sensed area, the reliability requirement, or network lifetime prolongation. It is necessary that the proposed protocols for WSNs are scalable for these large number of sensor nodes in order to achieve their tasks efficiently. \item \textbf{Reliability:} Many applications require high quality of services, connectivity, routing, data aggregation, etc. These applications need to deploy a large number of inexpensive sensor nodes so as to satisfy their requirements. This large number of the sensor nodes may be prone to failure and this will affect the quality of service provided by the application. However, it is important to build mechanisms inside the protocols so as to avoid the failure of some sensor nodes during the network operation and to increase the robustness of the proposed protocol in WSNs. @@ -193,7 +193,7 @@ The communication range of signals can be attenuated or faded during the signal \item \textbf{Data Management:} %It represents one of the challenges that contributes in depleting the energy of the sensor nodes in WSNs. -The main task of a WSN after deploying the sensor nodes in the target environment that need to be monitored, is to collect the sensed data from this physical environment and then transmit it to the base station. Since there are many sensor nodes in WSN and since every sensor node want to transmit its sensed data to the base station; there is a large amount of data that need to be managed, processed, and routed to the sink. Obviously, a suitable data management is required to minimize the corresponding energy consumption. +The main task of a WSN after deploying the sensor nodes in the target environment that need to be monitored, is to collect the sensed data from this physical environment and then transmit it to the base station. Since there are many sensor nodes in WSN and since each sensor node wants to transmit its sensed data to the base station; there is a large amount of data that need to be managed, processed, and routed to the sink. Obviously, a suitable data management is required to minimize the corresponding energy consumption. \item \textbf{Security:} The sensitivity of the information collected by WSNs represents the final challenge that should be faced in WSNs. An information is susceptible to malicious intrusions and hacker attacks. As a consequence, it is necessary to provide energy efficient schemes to protect this information during the operation of WSNs. @@ -268,7 +268,7 @@ The majority of synchronous schemes work in periodic (cyclic) way by preparing t %On the other hand, the aperiodic schemes do not apply the periodic schedule. \begin{enumerate} [(A)] -\item The periodic wakeup scheduling schemes work either in slotted and unslotted way, where the period is divided into equal-length slots in the slotted schemes. The major challenge in periodic wakeup scheduling is to select and activate the best time interval(s) for a period so that an active wireless sensor node performs the communication (sending and receiving). This is from point of view of wireless sensor node, whilst from the standpoint of the WSN, choosing the time intervals through the wireless sensor nodes to satisfy a certain performance factor seems to be hard task. This level of performance can be carried out with the cooperation among the sensor nodes in WSN to produce the wake-up schedule. The periodic wakeup scheduling schemes are classified into five groups based on the degree of a cooperation~\cite{ref57}: +\item The periodic wakeup scheduling schemes work either in slotted and unslotted way, where the period is divided into equal-length slots in the slotted schemes. The major challenge in periodic wakeup scheduling is to select and activate the best time interval(s) for a period so that an active wireless sensor node performs the communication (sending and receiving). This is from the point of view of wireless sensor node, whilst from the standpoint of the WSN, choosing the time intervals through the wireless sensor nodes to satisfy a certain performance factor seems to be hard task. This level of performance can be carried out with the cooperation among the sensor nodes in WSN to produce the wake-up schedule. The periodic wakeup scheduling schemes are classified into five groups based on the degree of a cooperation~\cite{ref57}: \begin{enumerate} [(i)] \item Neighbor-coordinated is a scheme in which a wireless sensor node generates its own wake-up schedule taking into consideration the wake-up schedules of its neighbor sensor nodes. @@ -288,7 +288,7 @@ The majority of synchronous schemes work in periodic (cyclic) way by preparing t \item \textbf{Asynchronous Schemes:} %The time among the wireless sensor nodes does not need synchronization. -The wireless sensor node wakes up to send packets without taking into account whether the receiving sensor nodes are waked up and ready to receive. These schemes do not need time synchronization which consumes energy~\cite{ref74}. They do not need to exploit the limited resources (processing, memory, and radio) of the sensor nodes because there are no shared wake up schedules to be exchanged or saved in the memory. Therefore, exchanging the packets among the wireless sensor nodes, which are not aware of each other's wake-up schedules, is a major challenge in asynchronous schemes. These schemes can been categorized into three groups~\cite{ref57}: +A wireless sensor node wakes up to send packets without taking into account whether the receiving sensor nodes are waked up and ready to receive. These schemes do not need time synchronization which consumes energy~\cite{ref74}. They do not need to exploit the limited resources (processing, memory, and radio) of the sensor nodes because there are no shared wake up schedules to be exchanged or saved in the memory. Therefore, exchanging the packets among the wireless sensor nodes, which are not aware of each other's wake-up schedules, is a major challenge in asynchronous schemes. These schemes can been categorized into three groups~\cite{ref57}: \begin{enumerate} [(A)] \item Transmitter-initiated: a special frame is sent by the transmitting sensor node to inform the receiving sensor node that it has a data frame to send. If the receiving sensor node is hearing the special frame during one of its wake up intervals, the receiving node waits for sending the data frame by sender to receive it. The major advantage of these schemes is the low memory and processing requirements whilst the major disadvantages are low-duty-cycle and the non-deterministic sleep latency. @@ -311,7 +311,7 @@ latency. \subsubsection{Topology Control Schemes} -\indent The topology control schemes deal with the redundancy in WSNs. The WSN is always deployed with high density and in a random way, where a large number of wireless sensor nodes are usually supposed to be thrown by the airplane over the area of interest. The purpose of deploying a dense WSN is to cope with the sensor failure during or after the WSN deployment and to maximize the network lifetime by means of exploiting the overlapping among the sensor nodes in the network. By putting the redundant sensor nodes into sleep mode, the idea is to benefit from the saved energy later. The major goal of topology control protocols is to dynamically adapt network topology based on requirements of application so as to minimize the number of active sensor nodes~\cite{ref56,ref22}. Many factors can be used to decide which sensor nodes should be turned on or off, and when. The topology control schemes have been classified into two categories~\cite{ref56}: +\indent The topology control schemes deal with the redundancy in WSNs. The WSN is sometimes deployed with high density and in a random way, where a large number of wireless sensor nodes are usually supposed to be thrown by the airplane over the area of interest. The purpose of deploying a dense WSN is to cope with the sensor failure during or after the WSN deployment and to maximize the network lifetime by means of exploiting the overlapping among the sensor nodes in the network. By putting the redundant sensor nodes into sleep mode, the idea is to benefit from the saved energy later. The major goal of topology control protocols is to dynamically adapt network topology based on requirements of application so as to minimize the number of active sensor nodes~\cite{ref56,ref22}. Many factors can be used to decide which sensor nodes should be turned on or off, and when. The topology control schemes have been classified into two categories~\cite{ref56}: \begin{enumerate} [(I)] \item \textbf{Location Driven Protocols:} Wireless sensor nodes are turned on or off based on their location; for example, Geographical Adaptive Fidelity (GAF) protocol~\cite{GAF}. @@ -445,7 +445,7 @@ where $R_u$ is a measure of the uncertainty in sensor detection, $\alpha = d(s_i \end{enumerate} -The coverage protocols proposed in this dissertation use the binary disc sensing model for each wireless sensor node in a WSN because it is widely used in the literature. Moreover, it is easy to formulate the linear programs with it, where as the probabilistic model is more complex and it is difficult to use it to create integer programs. +The coverage protocols proposed in this dissertation use the binary disc sensing model for each wireless sensor node in a WSN because it is widely used in the literature. Moreover, it is easy to formulate the linear programs with it, whereas the probabilistic model is more complex and it is difficult to use it to create integer programs. %The coverage protocols have proposed in this dissertation use the binary disc sensing model as a sensing coverage model for each wireless sensor node in WSN. @@ -458,22 +458,22 @@ The coverage protocols proposed in this dissertation use the binary disc sensing \indent Several design issues should be considered in order to produce solutions for the coverage problems in WSNs. These design issues can be classified into~\cite{ref103}: \begin{enumerate}[(i)] -\item $\textbf{Coverage Type}$ Is the WSN dedicated to monitor a whole area, to observe a set of targets, or to look for a breach of a barrier? +\item $\textbf{Coverage Type:}$ Is the WSN dedicated to monitor a whole area, to observe a set of targets, or to look for a barrier breach? -\item $\textbf{Deployment Method}$ refers to the way by which the wireless sensor nodes are deployed over the target sensing field in order to build the wireless sensor network. Generally, the sensor nodes can be placed either deterministically or randomly in the target sensing field~\cite{ref107}. The method of placement can be selected based on the type of sensors, application, and the environment. In the deterministic placement, the deployment can be achieved for a small number of sensor nodes and in friendly environment, whilst for a large number of sensor nodes or when the area of interest is inaccessible or hostile, a random placement is the choice. The sensor network can be either dense or sparse. On the one hand, the dense deployment is preferable when it is necessary to provide a security robustness in WSNs. On the other hand, the sparse deployment is used when the dense deployment is expensive or when the maximum coverage is performed by a low number of sensor nodes. +\item $\textbf{Deployment Method}$ refers to the way by which the wireless sensor nodes are deployed over the target sensing field in order to build the wireless sensor network. Generally, the sensor nodes can be placed either deterministically or randomly in the target sensing field~\cite{ref107}. The method of placement can be selected based on the type of sensors, application, and the environment. In the deterministic placement, the deployment can be achieved in a friendly environment with a small number of sensor nodes. The random placement is preferred for a large number of sensor nodes or when the area of interest is inaccessible or hostile. The sensor network can be either dense or sparse. On the one hand, the dense deployment is preferable when it is necessary to provide a security robustness in WSNs. On the other hand, the sparse deployment is used when the dense deployment is expensive or when the maximum coverage is performed by a low number of sensor nodes. -\item $\textbf{Coverage Degree}$ refers to how many sensor nodes are required to cover a target or an area. A point in the sensing field is said to be K-coverage by at least K sensor nodes. Some applications need a high reliability to achieve their tasks. Therefore, the sensing field is deployed densely so as to perform a K-coverage for this field. The simple coverage problem consists of a coverage degree equal to one (i.e., K=1), where every point in the sensing field is covered by at least one sensor. +\item $\textbf{Coverage Degree}$ refers to how many sensor nodes are required to cover a target or an area. A point in the sensing field is said to be K-coverage if it is covered by at least K sensor nodes. Some applications need a high reliability to achieve their tasks. Therefore, the sensing field is deployed densely so as to perform a K-coverage for this field. The simple coverage problem consists of a coverage degree equal to one (i.e., K=1), where every point in the sensing field is covered by at least one sensor. \item $\textbf{Coverage Ratio}$ is the percentage of the sensing field that fulfills the coverage degree of the application. If all the points in the sensing field are covered, the coverage ratio is $100\%$ and it can be called a complete coverage. Otherwise, it is said as partial coverage. -\item $\textbf{Network Connectivity}$ ensures the existence of a path from any sensor node in WSN to the sink. A connected WSN ensures the sending of the sending the sensed data from one sensor node to another sensor node directly toward the sink. +\item $\textbf{Network Connectivity}$ ensures the existence of a path from any sensor node in WSN to the sink. A connected WSN ensures the sending of the sensed data from one sensor node to another sensor node directly toward the sink. %It is necessary to consider the communication range of wireless sensor node is at least twice that of the sensing range ($R_c \geqslant 2R_s$) so as to imply connectivity among the sensor nodes during covering the sensing field~\cite{ref108}. -Activity based Scheduling schedules the activation and deactivation of sensor nodes during the network lifetime. +%Activity based Scheduling schedules the activation and deactivation of sensor nodes during the network lifetime. \item $\textbf{Activity based Scheduling}$ schedules the activation and deactivation of sensor nodes during the network lifetime. The basic objective is to decide which sensors are in what states (active or sleeping mode) and for how long, so that the application coverage requirement can be guaranteed and the network lifetime can be prolonged. Various centralized, distributed, and localized approaches have been proposed for activity scheduling. In distributed algorithms, each node in the network autonomously makes decisions on whether to turn on or turn off itself, using only local neighbor information. In centralized algorithms, a central controller (a node or base station) informs every sensor of the time intervals to be activated. -This dissertation deals with activity based scheduling to ensure the best coverage. +\textbf{This dissertation deals with activity based scheduling to ensure the best coverage}. \end{enumerate} @@ -485,7 +485,7 @@ In order to model the energy consumption, four states for a sensor node are used \begin{enumerate}[(i)] -\item Computation: processing needed for executing any algorithm inside the sensor node. The processing that is required to physical communication and networking protocols is included in reception and transmission. +\item Computation: processing needed to execute any algorithm inside the sensor node. The processing that is required to physical communication and networking protocols is included in reception and transmission. \item Transmission: processing for address determination, packetization, encoding, framing, and maybe queuing; supply for the baseband and RF circuitry. @@ -514,7 +514,7 @@ In order to model the energy consumption, four states for a sensor node are used \label{RDM} \end{figure} -\indent In this model, the radio consumes energy to execute the transmitter and the power amplifier. The receiver circuitry consumes energy to run the radio electronics, as described in figure~\ref{RDM}. The channel model can be either free space ($d^2$ power loss) or multipath fading ($d^4$ power loss), based on the distance between the transmitter and receiver. This power loss can be controlled by setting the power amplifier so that if the distance is less than a threshold ($d_0$), the free space ($\varepsilon_{fs}$) model is used (i.e., $\varepsilon_{amp}$ = $\varepsilon_{fs}$). Otherwise, the multipath ($\varepsilon_{mp}$) model is used (i.e., $\varepsilon_{amp}$ = $\varepsilon_{mp}$). Therefore, to transmit a K-bit packet across a distance d, the radio expends +\indent In this model, the radio consumes energy to execute the transmitter and the power amplifier. The receiver circuitry consumes energy to run the radio electronics, as described in figure~\ref{RDM}. The channel model can be either free space ($d^2$ power loss) or multipath fading ($d^4$ power loss), based on the distance between the transmitter and receiver. This power loss can be controlled by setting the power amplifier so that if the distance is less than a threshold ($d_0$), the free space ($\varepsilon_{fs}$) model is used (i.e., $\varepsilon_{amp}$ = $\varepsilon_{fs}$). Otherwise, the multipath ($\varepsilon_{mp}$) model is used (i.e., $\varepsilon_{amp}$ = $\varepsilon_{mp}$). Therefore, to transmit a K-bit packet across a distance d, the radio is \begin{equation} @@ -526,17 +526,17 @@ E_{tx}\left(K,d \right) = \left \{ \label{eq3-ch1} \end{equation} -while to receive a K-bit packet, the radio expends +while to receive a K-bit packet, the radio is \begin{equation} E_{rx}\left(k,d \right) = \emph{ KE_{elec} }. \label{eq4-ch1} \end{equation} -\noindent The typical parameters are set as: $E_{elec}$ = 50 nJ/bit, $\varepsilon_{fs}$ = 10 pJ/bit/$m^2$, $\varepsilon_{mp}$ = 0.0013 pJ/bit/$m^4$. In addition, the energy for data aggregation is set as $E_{DA}$ = 5 nJ/bit. +\noindent The typical parameters are set as: $E_{elec}$ = 50 nJ/bit, $\varepsilon_{fs}$ = 10 pJ/bit/$m^2$, $\varepsilon_{mp}$ = 0.0013 pJ/bit/$m^4$. In addition, the energy for data aggregation is set to $E_{DA}$ = 5 nJ/bit. \indent The radio energy dissipation model considers only the energy consumed by the communication part of the sensor node. However, in order to achieve a more accurate model, it is necessary to take into account the energy consumed by other parts inside the sensor node such as computation and sensing units. -In this dissertation, we developed another energy consumption model that based on \cite{ref112}. +\textbf{In this dissertation, we developed another energy consumption model that based on \cite{ref112}}. %\subsection{Our Energy Consumption Model:} %\label{ch1:sec9:subsec2} -- 2.39.5 From fc2e307dd138aca4c04532a3e6bed0aadff80225 Mon Sep 17 00:00:00 2001 From: ali Date: Tue, 5 May 2015 12:20:34 +0200 Subject: [PATCH 07/16] Update by Ali --- CONCLUSION.tex | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/CONCLUSION.tex b/CONCLUSION.tex index 9a9606a..0bc5569 100644 --- a/CONCLUSION.tex +++ b/CONCLUSION.tex @@ -7,39 +7,41 @@ \section{Conclusion} -In this dissertation, we have concentrated on proposing a distributed optimization protocols so as to prolong the lifetime of wireless sensor networks. We have addressed the problem of the area coverage and the lifetime optimization in wireless sensor networks, where the ultimate goal is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area (or region) of interest. +In this dissertation, we have concentrated on proposing a distributed optimization protocols so as to prolong the lifetime of wireless sensor networks. We have addressed the problem of the area coverage and the lifetime optimization in wireless sensor networks. The ultimate goal is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area of interest. The first part of the dissertation has presented the scientific background including WSNs, brief survey of related works, and evaluation tools as well as optimization solvers. -In chapter 1, We have began with a general overview on wireless sensor networks. We have described various concepts, mechanisms, types, applications, and challenges in WSNs. We have presented several energy-efficient techniques so as to improve the network lifetime of WSNs. The coverage problem, the network lifetime, and the energy consumption modeling in WSNs have explained. A brief survey about coverage algorithms in literature is achieved in chapter 2. -We have classified those works into centralized and distributed algorithms. We have given a brief comparison of the main characteristics of each approach. This part finally included in chapter 3, a comparative study of different evaluation tools dedicated to WSNs. In addition, we have illustrated a various commercial and free optimization solvers considering the main features of each one. +In chapter 1, We started with a general overview on wireless sensor networks. We have described various concepts, mechanisms, types, applications, and challenges in WSNs. Several energy-efficient techniques so as to improve the network lifetime of WSNs have been presented. The coverage problem, the network lifetime, and the energy consumption modeling in WSNs have been explained. A brief survey about literature on coverage algorithms is achieved in chapter 2. +We have classified those works into centralized and distributed algorithms. We have given a brief comparison of the main characteristics of each approach. Finally we have included in chapter 3 a comparative study of different evaluation tools dedicated to WSNs. In addition, we have illustrated various commercial and free optimization solvers considering the main features of each one. -In the second part of the dissertation, We have designed a new three different optimization protocols, which schedules nodes’ activities (wake up and sleep stages) with the objective of maintaining a good coverage ratio while maximizing the network lifetime in WSNs. This part proposes a two-step approaches. Firstly, the field of sensing is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of the proposed optimization protocols is applied in each subregion in a distributed parallel way to optimize the coverage and lifetime performances. The proposed protocols combine two efficient mechanisms: network leader election and sensor activity scheduling, where the challenges include how to select the most efficient leader in each subregion and the best +In the second part of the dissertation, We have designed three new different optimization protocols, which schedules nodes’ activities (wake up and sleep stages) with the objective of maintaining a good coverage ratio while maximizing the network lifetime. We propose two-step approaches. Firstly, the field of sensing is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of the proposed optimization protocols is applied in each subregion in a distributed parallel way to optimize both coverage and lifetime performances. The proposed protocols combine two efficient mechanisms: network leader election and sensor activity scheduling, where the challenges include how to select the most efficient leader in each subregion, the best representative active nodes that will optimize the network lifetime while taking the responsibility of covering the corresponding subregion. -In chapter 4, we have proposed an optimization protocol, called Distributed Lifetime Coverage Optimization protocol (DiLCO), which optimizes the coverage and the lifetime of a wireless sensor network. DiLCO protocol is distributed in each sensor node in the subregions of the sensing field. It has been implemented in each subregion simultaneously and Independently. The proposed DiLCO protocol is a periodic protocol where each period consists of 4 phases: information exchange, leader election, decision, and sensing. The sensor nodes collaborate in each subregion to elect the leader. The leader applies activity scheduling based optimization so as to provides only one optimal set of active sensor nodes that takes the mission of sensing during this period. The performance of our DiLCO protocol has been evaluated by a series of simulations. We have presented a comparison between our proposed protocol and other two existing and known in the literature, DESK and GAF. The experimental results have validated our protocol and showed their efficiency in the optimization of the coverage and the lifetime compared to existing methods. +In chapter 4, we have proposed an optimization protocol called Distributed Lifetime Coverage Optimization protocol (DiLCO), which optimizes the coverage and the lifetime of a wireless sensor network. DiLCO protocol is distributed in each sensor node in the subregions of the sensing field. It is implemented in each subregion simultaneously and independently. The proposed DiLCO protocol is a periodic protocol where each period consists of 4 phases: information exchange, leader election, decision, and sensing. The sensor nodes collaborate in each subregion to elect the leader. The leader applies activity scheduling based optimization in order to provide only one optimal set of active sensor nodes that takes the mission of sensing during this period. The performance of our DiLCO protocol has been evaluated by a series of simulations. We have presented a comparison between our proposed protocol and two other existing protocols known in the literature: DESK and GAF. The experimental results have validated our protocol and showed its efficiency in the optimization of the coverage and the lifetime compared to the two references. -Next, we propose a method called Multiround Distributed Lifetime Coverage Optimization protocol (MuDiLCO) in chapter 5, to maintain the coverage and to improve the lifetime in wireless sensor networks. MuDiLCO protocol is an extension of the DiLCO protocol introduced in chapter 4. In MuDiLCO, the protocol has implemented activity scheduling based optimization in order provides a multiple set of active sensor nodes for several rounds in the sensing phase. We have introduced an improved coverage optimization model that make a multiround optimization, whilst it was a single round optimization in DiLCO. We have conducted several sets of simulations comparing the proposed MuDiLCO protocol for different number of rounds as well as with other existing coverage methods like DESK and GAF. +Next, we propose in chapter 5 a method called Multiround Distributed Lifetime Coverage Optimization protocol (MuDiLCO), which is an extension of the DiLCO protocol introduced in chapter 4. MuDiLCO implemented an activity scheduling based optimization in order to provide multiple sets of active sensor nodes, for several rounds in the sensing phase. We have thus introduced an improved coverage optimization model that make a multiround optimization, whilst it was a single round optimization in DiLCO. We have conducted many simulations comparing the proposed MuDiLCO protocol for different number of rounds, as well as with DiLCO, DESK, and GAF. -In chapter 6, We have proposed an approach called Perimeter-based Coverage Optimization protocol (PeCO) in order to optimize the lifetime coverage, so that it provides activity scheduling which ensures sensing coverage as long as possible. PeCO protocol is distributed among sensor nodes in each subregion. The novelty of our approach lies essentially in the formulation of a new mathematical optimization model based on the perimeter coverage level to schedule sensors’ activities. The leader provides one schedule during the current period by executing the new integer program during the decision phase. The extensive simulation experiments have demonstrated that PeCO can offer longer lifetime coverage for WSNs in comparison with some other protocols. +In chapter 6, we have proposed an approach called Perimeter-based Coverage Optimization protocol (PeCO) in order to optimize the lifetime coverage, so that it provides activity scheduling which ensures sensing coverage as long as possible. Like DiLCO and MuDiLCO, PeCO protocol is distributed among sensor nodes in each subregion. The novelty of our approach, in comparison with DiLCO and MuDiLCO, lies essentially in the formulation of a new mathematical optimization model based on the perimeter coverage level to schedule sensors’ activities. A leader provides one schedule during the current period by executing the new integer program during the decision phase. The extensive simulation experiments have demonstrated that PeCO can offer longer lifetime coverage for WSNs. -Finally, we outline some interesting issues that we will consider in our perspectives which are discussed in more detail next. +Finally, we outlined some interesting issues that will be considered in our perspectives which are discussed in more detail next. \section{Perspectives} -In this dissertation, we have focused on the lifetime area coverage optimization problem and we are interested only in energy-efficient, distributed and parallel protocol. Various scenarios might need to be taken into consideration such as fault-tolerance, k-coverage, $\alpha$-coverage, adjustable sensor’s sensing range network, heterogeneous network, mobility, etc. In the future, we will concentrate on the following work: +In this dissertation, we have focused on the lifetime area coverage optimization problem and we have interested only in energy-efficient distributed protocols. Various scenarios might need to be taken into consideration such as fault-tolerance, k-coverage, $\alpha$-coverage, adjustable sensor’s sensing range network, heterogeneous network, mobility, etc. In the future, we will concentrate on the following work: -In chapter 4, We have studied the impact of the number of subregions chosen to subdivide the area of interest, considering different network sizes. The optimal number of subregions will be investigated in the future. We also plan to study and propose a coverage protocol, which computes all active sensor schedules in one time, using optimization methods such as swarms optimization or evolutionary algorithms. The round will still consist of 4 phases, but the decision phase will compute the schedules for several sensing phases which, aggregated together, define a kind of meta-sensing phase. The computation of all cover sets in one time is far more difficult, but will reduce the communication overhead. +In chapter 4, We have studied the impact of the number of subregions chosen to subdivide the area of interest, considering different network sizes. The optimal number of subregions will be investigated in the future. We also plan to study and propose a coverage protocol, which computes all active sensor schedules in one time, using optimization methods such as particle swarm optimization or evolutionary algorithms. A period will still consist of 4 phases, but the decision phase will compute the schedules for several sensing rounds which, aggregated together, define a kind of meta-sensing round. The computation of all cover sets in one step is far more difficult, but will reduce the communication overhead. -In chapter 5, we plan to design and propose a heterogeneous integrated optimization protocol in WSNs. This protocol integrates three energy-efficient (coverage, routing and data aggregation) protocols so as to extend the network lifetime in WSNs. The sensing, routing, and aggregation jobs are also challenges in WSNs. This integrated optimization protocol will be executed by each cluster head in the wireless sensor network. The cluster head will be selected in a distributed way and based on local information. +We also plan to design and propose a heterogeneous integrated optimization protocol in WSNs. This protocol would integrate three energy-efficient (coverage, routing and data aggregation) protocols so as to extend the network lifetime in WSNs. The sensing, routing, and aggregation jobs are also challenges in WSNs. This integrated optimization protocol will be executed by each cluster head, a leader node in our protocols, in the wireless sensor network. The cluster head will be selected in a distributed way and based on local information. -In chapter 6, We plan to extend our framework so that the schedules are planned for multiple sensing periods. We also want to improve our integer program to take into account heterogeneous sensors from both energy and node characteristics point of views. Finally, it would be interesting to implement our protocol using a sensor-testbed to evaluate it in real world applications. +We plan to extend our PeCO protocol so that the schedules are planned for multiple sensing periods. We also want to improve our integer program to take into account heterogeneous sensors from both energy and node characteristics point of views. + +Finally, it would be interesting to implement our protocols using a sensor-testbed to evaluate it in real world applications. -- 2.39.5 From 6d401cb266bc54619b552e898e3717db5c4b4f59 Mon Sep 17 00:00:00 2001 From: ali Date: Tue, 5 May 2015 19:54:29 +0200 Subject: [PATCH 08/16] Update for French Abstract by Ali. --- Resume.tex | 19 +++++++------------ 1 file changed, 7 insertions(+), 12 deletions(-) diff --git a/Resume.tex b/Resume.tex index 15fa445..9600b1f 100644 --- a/Resume.tex +++ b/Resume.tex @@ -16,22 +16,17 @@ \emph{ \begin{center} \large Encadrants: Raphaël Couturier, Karine Deschinkel, and Michel Salomon \end{center}} -Réseaux de capteurs sans fil ont récemment reçu beaucoup d'attention de la recherche en raison de leur large gamme d'applications potentielles. Beaucoup de caractéristiques importantes sont fournis par les réseaux de capteurs qui les rendent différent des autres réseaux ad-hoc sans fil. Ces caractéristiques sont imposées beaucoup de limitations sur les réseaux de capteurs qui mèneraient à plusieurs défis dans le réseau. Ces défis pourraient inclure la couverture, contrôle de topologie, routage, la fusion de données, la sécurité, et bien d'autres. L'un des principaux défis de la recherche rencontrés dans les réseaux de capteurs sans fil est de préserver effectivement et en permanence la couverture d'une zone d'intérêt à surveiller, tout en empêchant simultanément autant que possible une défaillance du réseau en raison de nœuds de batterie appauvri. +Les réseaux de capteurs sans fil ont suscité beaucoup d'intérêt dans le domaine de la recherche au cours des dernières années en raison de leur large gamme d'applications potentielles. Ils fournissent de nombreuses caractéristiques importantes qui les rendent différents des autres réseaux ad-hoc sans fil. Néanmoins ces caractéristiques imposent beaucoup de limitations susceptibles de créer plusieurs défis dans le domaine des réseaux. Ces défis pourraient inclure la couverture, le contrôle de topologie, le routage, la fusion de données, la sécurité, et bien d'autres. L'une des principales problématiques de recherche étudiée dans les réseaux de capteurs sans fil est la préservation de la couverture d'une zone à surveiller d'une manière permanente et efficace, tout en empêchant autant que possible le dysfonctionnement du réseau en raison de déchargement de batterie de certains n\oe uds. -Dans cette thèse, nous nous concentrons fortement sur le problème de la zone de couverture, l'efficacité énergétique est également l'exigence avant tout. Nous avons examiné les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Les protocoles proposés distribués d'optimisation (y compris les algorithmes, les modèles, et la résolution des programmes entiers) doivent être protocoles économes en énergie. Adresser ce problème, cette thèse propose des approches en deux étapes. Tout d'abord, le champ de détection est divisée en plus petites sous-régions en utilisant le concept de la méthode de diviser pour régner. Deuxièmement, l'un de nos protocoles d'optimisation distribués proposées est distribuée et appliquée sur les nœuds de capteurs dans chaque sous-région afin d'optimiser la couverture et les performances de durée de vie. Dans cette thèse, trois protocoles d'optimisation de couverture sont proposés. Ces protocoles combinent deux techniques efficaces: élection du chef pour chaque sous-région, suivis par une planification fondée sur l'optimisation des décisions de planification d'activité du capteur pour chaque sous-région. +Dans cette thèse, nous nous sommes intéressés au problème de la zone de couverture ainsi qu'à l'efficacité énergétique qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Les protocoles d'optimisation distribués proposés (y compris les algorithmes, les modèles et la résolution des programmes entiers) doivent être efficaces en terme d'énergie. Pour résoudre ce problème, nous avons proposé de nouvelles approches en deux phases. Dans un premier temps, le champ de surveillance est divisé en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Ensuite, l'un de nos protocoles d'optimisation distribués est appliqué sur les n\oe uds de capteurs dans chaque sous-régions afin d'optimiser la couverture et la durée de vie du réseau. Dans cette thèse, nous avons proposé trois protocoles distribués pour l'optimisation de la couverture. Ces protocoles permettent de combiner deux techniques efficaces: une élection de leader pour chaque sous-région, suivie par un processus d'optimisation de l'ordonnancement d'activité de décisions des capteurs pour chaque sous-région. -Premièrement, nous proposons un protocole appelé Optimisation de couverture à vie (Distributed DILCO). Dans ce protocole, la durée de vie est divisée en périodes. Chaque période se compose de quatre phases: échange d'informations, leader électorales, de décision et de détection. Le processus de décision est -effectuée par un nœud leader, qui résout un programme entier afin de fournir un seul ensemble de nœuds de capteurs actifs de couverture pour assurer une couverture pendant la phase de détection de la période actuelle. +Premièrement, nous avons proposé un protocole appelé optimisation distribuée de la durée de vie de la couverture (DILCO). Dans ce protocole, la durée de vie est divisée en périodes. Chaque période se compose de quatre phases: échange d'informations, élection de leader, décision et surveillance. Le processus de décision est effectué par le n\oe ud leader, qui résout un programme entier permettant de définir un seul ensemble de n\oe uds de capteurs actifs pour assurer la couverture durant une période. -Ensuite, nous abordons le problème d'une optimisation des passages répétés problème de la couverture de la zone dans les réseaux de capteurs. Le passages répétés Optimisation de couverture à vie (Distributed MuDiLCO) protocole est suggéré afin d'étudier la possibilité de fournir de multiples ensembles de couverture des capteurs pour la phase de détection. Protocole MuDiLCO travaille également en périodes pendant lesquelles ensembles de nœuds de capteurs sont programmés pour rester actif pour un certain nombre de tours pendant la phase de détection, pour assurer une couverture de manière à maximiser la durée de vie de réseaux de capteur sans fil. Le processus de décision est toujours effectuée par un nœud leader, qui résout un programme entier pour produire le meilleur représentant établit à être utilisé pendant les tours de la phase de détection. +Ensuite, nous avons étudié le problème de l'optimisation multi-ronde de la zone de couverture dans un réseau de capteurs sans fil. Nous avons proposé le protocole d'optimisation multi-ronde distribué de la durée de vie de couverture (MuDiLCO) pour étudier la possibilité de fournir plusieurs ensembles de n\oe uds de capteurs de couverture pour la phase de surveillance. Ce protocole travaille également en périodes pendant lesquelles les ensembles de capteurs sont programmés pour rester actifs pour un certain nombre de rondes durant la phase de surveillance, pour assurer la couverture et maximiser la durée de vie du réseau. Le processus de décision est toujours effectué par le n\oe ud leader qui résout un programme entier pour définir un meilleur ensemble de capteurs à être utilisé pendant les rondes de la phase de surveillance. +Enfin, nous avons proposé le protocole d'optimisation de la couverture basé sur le périmètre (PeCO) qui est aussi un protocole distribué sur les n\oe uds de capteurs dans chaque sous-région. Notre contribution dans ce protocole consiste essentiellement dans la proposition d'un nouveau modèle mathématique de l'optimisation basé sur le périmètre de couverture pour l'ordonnancement de l'activité des capteurs. Un nouveau programme entier du modèle de couverture est résolu par le leader durant la phase de décision pour définir un ensemble de capteurs de couverture pour la phase de surveillance. +Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNET++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. -Enfin et surtout,, nous proposons une couverture Optimization (Peco) protocole basé périmètre qui est également réparti entre les nœuds de capteurs dans chaque nouveauté subregion.The de notre approche réside essentiellement dans la formulation d'un nouveau modèle d'optimisation mathématique basée sur le niveau de couverture de périmètre pour planifier les activités de capteurs. Un nouveau modèle de couverture du programme entier est résolu par le leader pendant la phase de décision de façon à fournir un seul ensemble de capteurs de couverture pour la phase de détection. - - -Simulations approfondies sont menées en utilisant la simulation à événements discrets OMNeT++ pour valider l'efficacité de chacun de nos protocoles proposés. Nous nous référons à la características capteur de méduse II de la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles sont fiables pour augmenter la couverture à vie de réseaux de capteur sans fil et améliorent les performances. - - -\textbf{MOTS-CLÉS:} Réseaux sans fil, les réseaux de capteurs, Zone de couverture, Durée de vie du réseau, optimisation, la planification, algorithmes distribués, Algorithmes centralisée, Robustesse, connectivité, l'efficacité énergétique, l'énergie réseau hétérogène, homogène réseau. +\textbf{MOTS-CLÉS:} Réseaux sans fil, Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation, Ordonnancement, Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Algorithmes parallèles, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes. -- 2.39.5 From 508b0afd303ff3341d65be0960746229924e9863 Mon Sep 17 00:00:00 2001 From: ali Date: Wed, 6 May 2015 18:44:28 +0200 Subject: [PATCH 09/16] Update by Ali --- ACRONYMS.tex | 2 +- Abstruct.tex | 8 ++++---- CHAPITRE_03.tex | 2 +- CONCLUSION.tex | 10 +++++----- Resume.tex | 19 +++++++++++++------ Thesis.tex | 1 - entete.tex | 9 +++++---- 7 files changed, 29 insertions(+), 22 deletions(-) diff --git a/ACRONYMS.tex b/ACRONYMS.tex index f30ab04..a95c3df 100644 --- a/ACRONYMS.tex +++ b/ACRONYMS.tex @@ -53,7 +53,7 @@ \item[GUI] Graphical User Interface \item[NED] NEtwork Description \item[ns-2] Network Simulator-2 -\item[OPNET] Optimized Network Engineering tool +\item[OPNET] Optimized Network Engineering Tool \item[GloMoSim] Global Mobile System Simulator \item[SENSE] Sensor Network Simulator and Emulator \item[GTSNetS] Georgia Tech Sensor Network Simulator diff --git a/Abstruct.tex b/Abstruct.tex index 103d2ba..cfa6033 100644 --- a/Abstruct.tex +++ b/Abstruct.tex @@ -19,13 +19,13 @@ -Wireless sensor networks (WSNs) have recently received a great deal of research attention due to their wide range of potential applications. Many important characteristics are provided by the WSNs which make them different from other wireless ad-hoc networks. These characteristics are imposed lots of limitations on the WSNs that would lead to several challenges in the network. These challenges might include coverage, topology control, routing, data fusion, security, and many others. One of the main research challenges faced in wireless sensor networks is to preserve continuously and effectively the coverage of an area of interest to be monitored, while simultaneously preventing as much as possible a network failure due to battery-depleted nodes. +Wireless sensor networks (WSNs) have recently received a great deal of research attention due to their wide range of potential applications. Many important characteristics provided by the WSNs make them different from other wireless ad-hoc networks. Furthermore, these characteristics impose lots of limitations that lead to several challenges in the network. These challenges include coverage, topology control, routing, data fusion, security, and many others. One of the main research challenges faced in wireless sensor networks is to preserve continuously and effectively the coverage of an area of interest to be monitored, while simultaneously preventing as much as possible a network failure due to battery-depleted nodes. In this dissertation, we highly focus on the area coverage problem, energy-efficiency is also the foremost requirement. We have considered distributed optimization protocols with the ultimate objective of prolonging the network lifetime. The proposed distributed optimization protocols (including algorithms, models, and solving integer programs) should be energy-efficient protocols. To address this problem, this dissertation proposes two-step approaches. Firstly, the sensing field is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of our proposed distributed optimization protocols is distributed and applied on the sensor nodes in each subregion so as to optimize the coverage and the lifetime performances. In this dissertation, three coverage optimization protocols are proposed. These protocols combine two efficient techniques: leader election for each subregion, followed by an optimization-based planning of sensor activity scheduling decisions for each subregion. -First, we propose a protocol called Distributed Lifetime Coverage Optimization (DILCO). In this protocol, the lifetime is divided into periods. Each period consists of 4 phases: information exchange, leader election, decision, and sensing. The decision process is +First, we propose a protocol called Distributed Lifetime Coverage Optimization (DiLCO). In this protocol, the lifetime is divided into periods. Each period consists of 4 phases: information exchange, leader election, decision, and sensing. The decision process is carried out by a leader node, which solves an integer program in order to provide only one cover set of active sensor nodes to ensure coverage during the sensing phase of the current period. Then we address the problem of a multiround optimization of the area coverage problem in WSNs. The Multiround Distributed Lifetime Coverage Optimization (MuDiLCO) protocol is suggested so as to study the possibility of providing multiple cover sets of sensors for the sensing phase. MuDiLCO protocol also works in periods @@ -34,11 +34,11 @@ during which sets of sensor nodes are scheduled to remain active for a number of Last but not least, we propose a Perimeter-based Coverage Optimization (PeCO) protocol which is also distributed among sensor nodes in each subregion.The novelty of our approach lies essentially in the formulation of a new -mathematical optimization model based on the perimeter coverage level to schedule sensors' activities. A new integer program coverage model is solved by the leader during the decision phase so as to provide only one cover set of sensors for the sensing phase. +mathematical optimization model based on a perimeter coverage level to schedule sensors' activities, whereas we used primary points coverage model in the two previous models. A new integer program coverage model is solved by the leader during the decision phase so as to provide only one cover set of sensors for the sensing phase. Extensive simulations are conducted using the discrete event simulator OMNET++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase the WSN lifetime and provide improved coverage performance. -\textbf{KEY WORDS:} Wireless Networks, Wireless Sensor Networks, Area Coverage, Network Lifetime, Optimization, Scheduling, Distributed Algorithms, Centralized Algorithms, Robustness, Connectivity, Parallel Algorithms, Energy-efficiency, Heterogeneous Energy Network, Homogeneous Network. +\textbf{KEY WORDS:} Wireless Sensor Networks, Area Coverage, Network Lifetime, Distributed Optimization, Scheduling. diff --git a/CHAPITRE_03.tex b/CHAPITRE_03.tex index a4dd3fa..93b3b3b 100644 --- a/CHAPITRE_03.tex +++ b/CHAPITRE_03.tex @@ -112,7 +112,7 @@ OMNeT++ (Objective Modular Network Testbed) is an open-source, free, discrete-ev \item \textbf{OPNET:} -OPNET (Optimized Network Engineering tool)~\cite{ref192,ref200,ref201} is the first commercial simulation tool developed in 1987 for communication networks. It is a discrete event, object-oriented, general purpose network simulator, which is widely used in industry. It uses C and Java languages. It provides a comprehensive development environment for the specification, simulation, configuration, and performance analysis of the communication network. OPNET allows researchers to develop various models by means of a graphical interface. It provides different types of tools such as Probe Editor, Filter Tool, and Animation Viewer for data collection to model graph and animate the resulting output. Unlike ns-2, OPNET provides modeling for different sensor-specific hardware, such as physical-link transceivers and antennas. It includes sensor-specific models such as ad-hoc connectivity, mobility of nodes, node failure models, modeling of power-consumption, etc. OPNET is, a commercial simulator and the license is very expensive. This represents the main disadvantage of that simulator. +OPNET (Optimized Network Engineering Tool)~\cite{ref192,ref200,ref201} is the first commercial simulation tool developed in 1987 for communication networks. It is a discrete event, object-oriented, general purpose network simulator, which is widely used in industry. It uses C and Java languages. It provides a comprehensive development environment for the specification, simulation, configuration, and performance analysis of the communication network. OPNET allows researchers to develop various models by means of a graphical interface. It provides different types of tools such as Probe Editor, Filter Tool, and Animation Viewer for data collection to model graph and animate the resulting output. Unlike ns-2, OPNET provides modeling for different sensor-specific hardware, such as physical-link transceivers and antennas. It includes sensor-specific models such as ad-hoc connectivity, mobility of nodes, node failure models, modeling of power-consumption, etc. OPNET is, a commercial simulator and the license is very expensive. This represents the main disadvantage of that simulator. \item \textbf{GloMoSim:} diff --git a/CONCLUSION.tex b/CONCLUSION.tex index 0bc5569..6706da0 100644 --- a/CONCLUSION.tex +++ b/CONCLUSION.tex @@ -7,7 +7,7 @@ \section{Conclusion} -In this dissertation, we have concentrated on proposing a distributed optimization protocols so as to prolong the lifetime of wireless sensor networks. We have addressed the problem of the area coverage and the lifetime optimization in wireless sensor networks. The ultimate goal is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area of interest. +In this dissertation, we have concentrated on on the design of distributed optimization protocols so as to prolong the lifetime of wireless sensor networks. We have addressed the problem of the area coverage and the lifetime optimization in wireless sensor networks. The ultimate goal is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area of interest. The first part of the dissertation has presented the scientific background including WSNs, brief survey of related works, and evaluation tools as well as optimization solvers. @@ -19,7 +19,7 @@ representative active nodes that will optimize the network lifetime while taking -In chapter 4, we have proposed an optimization protocol called Distributed Lifetime Coverage Optimization protocol (DiLCO), which optimizes the coverage and the lifetime of a wireless sensor network. DiLCO protocol is distributed in each sensor node in the subregions of the sensing field. It is implemented in each subregion simultaneously and independently. The proposed DiLCO protocol is a periodic protocol where each period consists of 4 phases: information exchange, leader election, decision, and sensing. The sensor nodes collaborate in each subregion to elect the leader. The leader applies activity scheduling based optimization in order to provide only one optimal set of active sensor nodes that takes the mission of sensing during this period. The performance of our DiLCO protocol has been evaluated by a series of simulations. We have presented a comparison between our proposed protocol and two other existing protocols known in the literature: DESK and GAF. The experimental results have validated our protocol and showed its efficiency in the optimization of the coverage and the lifetime compared to the two references. +In chapter 4, we have proposed an optimization protocol called Distributed Lifetime Coverage Optimization protocol (DiLCO), which optimizes the coverage and the lifetime of a wireless sensor network. DiLCO protocol is distributed in each sensor node in the subregions of the sensing field. It is implemented in each subregion simultaneously and independently. The proposed DiLCO protocol is a periodic protocol where each period consists of 4 phases: information exchange, leader election, decision, and sensing. The sensor nodes collaborate in each subregion to elect the leader. The leader applies activity scheduling based optimization in order to provide only one optimal set of active sensor nodes that takes the mission of sensing during this period. The performance of our DiLCO protocol has been evaluated by a series of simulations. We have presented a comparison between our proposed protocol and two other existing protocols known in the literature: DESK and GAF. The experimental results have validated our protocol and showed its efficiency in the optimization of the coverage and the lifetime compared to the two benchmarking methods. Next, we propose in chapter 5 a method called Multiround Distributed Lifetime Coverage Optimization protocol (MuDiLCO), which is an extension of the DiLCO protocol introduced in chapter 4. MuDiLCO implemented an activity scheduling based optimization in order to provide multiple sets of active sensor nodes, for several rounds in the sensing phase. We have thus introduced an improved coverage optimization model that make a multiround optimization, whilst it was a single round optimization in DiLCO. We have conducted many simulations comparing the proposed MuDiLCO protocol for different number of rounds, as well as with DiLCO, DESK, and GAF. @@ -31,10 +31,10 @@ Finally, we outlined some interesting issues that will be considered in our pers \section{Perspectives} +In this dissertation, we have focused on the lifetime area coverage optimization problem and we were interested only in energy-efficient distributed protocols, considering static homogeneous sensor nodes. Several parameters, constraints, and requirements can have an important impact on the coverage performance in WSNs. +Thus, various scenarios parameters might need to be taken into consideration in the future, such as fault-tolerance, k-coverage, $\alpha$-coverage, adjustable sensor’s sensing range network, heterogeneous network, mobility, etc. -In this dissertation, we have focused on the lifetime area coverage optimization problem and we have interested only in energy-efficient distributed protocols. Various scenarios might need to be taken into consideration such as fault-tolerance, k-coverage, $\alpha$-coverage, adjustable sensor’s sensing range network, heterogeneous network, mobility, etc. In the future, we will concentrate on the following work: - -In chapter 4, We have studied the impact of the number of subregions chosen to subdivide the area of interest, considering different network sizes. The optimal number of subregions will be investigated in the future. We also plan to study and propose a coverage protocol, which computes all active sensor schedules in one time, using optimization methods such as particle swarm optimization or evolutionary algorithms. A period will still consist of 4 phases, but the decision phase will compute the schedules for several sensing rounds which, aggregated together, define a kind of meta-sensing round. The computation of all cover sets in one step is far more difficult, but will reduce the communication overhead. +In chapter 4, we have studied the impact of the number of subregions chosen to subdivide the area of interest, considering different network sizes. The optimal number of subregions will be investigated in the future. We also plan to study and propose a coverage protocol, which computes all active sensor schedules in one time, using optimization methods such as particle swarm optimization or evolutionary algorithms. A period will still consist of 4 phases, but the decision phase will compute the schedules for several sensing rounds which, aggregated together, define a kind of meta-sensing round. The computation of all cover sets in one step is far more difficult, but will reduce the communication overhead. We also plan to design and propose a heterogeneous integrated optimization protocol in WSNs. This protocol would integrate three energy-efficient (coverage, routing and data aggregation) protocols so as to extend the network lifetime in WSNs. The sensing, routing, and aggregation jobs are also challenges in WSNs. This integrated optimization protocol will be executed by each cluster head, a leader node in our protocols, in the wireless sensor network. The cluster head will be selected in a distributed way and based on local information. diff --git a/Resume.tex b/Resume.tex index 9600b1f..5dd0a97 100644 --- a/Resume.tex +++ b/Resume.tex @@ -9,24 +9,31 @@ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -\emph{ \begin{center} \Large Techniques d'Optimisation Couverture Distribuée pour Améliorer la Durée des Réseaux de Capteurs sans Fil \end{center}} +\emph{ \begin{center} \Large Techniques Distribuées d'Optimisation de la Couverture des Réseaux de Capteurs sans Fil pour Améliorer leur Durée de Vie \end{center}} %\emph{ \begin{center} \large By \end{center}} \emph{ \begin{center} \large Ali Kadhum Idrees \\ Université de Franche-Comt\'e, 2015 \end{center}} %\emph{ \begin{center} \large The University of Franche-Comt\'e, 2015 \end{center}} \emph{ \begin{center} \large Encadrants: Raphaël Couturier, Karine Deschinkel, and Michel Salomon \end{center}} -Les réseaux de capteurs sans fil ont suscité beaucoup d'intérêt dans le domaine de la recherche au cours des dernières années en raison de leur large gamme d'applications potentielles. Ils fournissent de nombreuses caractéristiques importantes qui les rendent différents des autres réseaux ad-hoc sans fil. Néanmoins ces caractéristiques imposent beaucoup de limitations susceptibles de créer plusieurs défis dans le domaine des réseaux. Ces défis pourraient inclure la couverture, le contrôle de topologie, le routage, la fusion de données, la sécurité, et bien d'autres. L'une des principales problématiques de recherche étudiée dans les réseaux de capteurs sans fil est la préservation de la couverture d'une zone à surveiller d'une manière permanente et efficace, tout en empêchant autant que possible le dysfonctionnement du réseau en raison de déchargement de batterie de certains n\oe uds. +Les réseaux de capteurs sans fil ont suscité beaucoup de travaux de recherche au cours des dernières années en raison de leur large gamme d'applications potentielles. Les caractéristiques des noeuds capteurs imposent des contraints enterme de consommation d'énergie et de capacité de traitement qui rendent caduque les protocoles des réseaux ad-hoc sans fil, avec de nombreux défis à résoudre. Parmi ces défis, on peut noter la préservation de la couverture, le contrôle de la topologie, le routage, la fusion de données, la sécurité, etc. La préservation de la couverture d'une région à surveiller, de manière permanente et efficace, tout en empêchant autant que possible un dysfonctionnement du réseau en raison du déchargement de la batterie de certains n\oe uds, est une des problématique de recherche majeures. -Dans cette thèse, nous nous sommes intéressés au problème de la zone de couverture ainsi qu'à l'efficacité énergétique qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Les protocoles d'optimisation distribués proposés (y compris les algorithmes, les modèles et la résolution des programmes entiers) doivent être efficaces en terme d'énergie. Pour résoudre ce problème, nous avons proposé de nouvelles approches en deux phases. Dans un premier temps, le champ de surveillance est divisé en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Ensuite, l'un de nos protocoles d'optimisation distribués est appliqué sur les n\oe uds de capteurs dans chaque sous-régions afin d'optimiser la couverture et la durée de vie du réseau. Dans cette thèse, nous avons proposé trois protocoles distribués pour l'optimisation de la couverture. Ces protocoles permettent de combiner deux techniques efficaces: une élection de leader pour chaque sous-région, suivie par un processus d'optimisation de l'ordonnancement d'activité de décisions des capteurs pour chaque sous-région. -Premièrement, nous avons proposé un protocole appelé optimisation distribuée de la durée de vie de la couverture (DILCO). Dans ce protocole, la durée de vie est divisée en périodes. Chaque période se compose de quatre phases: échange d'informations, élection de leader, décision et surveillance. Le processus de décision est effectué par le n\oe ud leader, qui résout un programme entier permettant de définir un seul ensemble de n\oe uds de capteurs actifs pour assurer la couverture durant une période. +Dans cette thèse, nous nous sommes intéressés au problème de la préservation de la couverture, ainsi qu'à l'efficatité qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie opérationnelle du réseau. Les protocoles proposés doivent être efficaces en terme de consommation énergétique induite par les calculs et les communications. Pour résoudre le problème, nous avons proposé des nouvelles approches en deux étapes. Dans un premier temps, la région à surveiller est divisée en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Dans un second temps, un de nos protocoles est exécuté par chacun des noeuds capteurs dans chaque sous-région, afin d'optimiser la couverture et la durée de vie du réseau. Nous proposons trois protocoles distribués qui combinent, chacun, deux techniques efficaces: l'élection d'un noeud leader dans chaque sous-région, suivie par la mise en oeuvre par celui-ci d'un processus de décision via l'optimisation de l'ordonnancement d'activité des noeuds capteurs de sa sous-région. + + + + + + + +Le premier protocole proposé est appelé DiLCO, pour Distributed Lifetime Coverage Optimization. Dans ce protocole, la durée de vie est divisée en périodes. Chaque période se compose de quatre phases: échange d'informations, élection de leader, décision et surveillance. Le processus de décision est effectué par le n\oe ud leader, qui résout un programme entier permettant de définir un seul ensemble de n\oe uds de capteurs actifs pour assurer la couverture durant une période. Ensuite, nous avons étudié le problème de l'optimisation multi-ronde de la zone de couverture dans un réseau de capteurs sans fil. Nous avons proposé le protocole d'optimisation multi-ronde distribué de la durée de vie de couverture (MuDiLCO) pour étudier la possibilité de fournir plusieurs ensembles de n\oe uds de capteurs de couverture pour la phase de surveillance. Ce protocole travaille également en périodes pendant lesquelles les ensembles de capteurs sont programmés pour rester actifs pour un certain nombre de rondes durant la phase de surveillance, pour assurer la couverture et maximiser la durée de vie du réseau. Le processus de décision est toujours effectué par le n\oe ud leader qui résout un programme entier pour définir un meilleur ensemble de capteurs à être utilisé pendant les rondes de la phase de surveillance. Enfin, nous avons proposé le protocole d'optimisation de la couverture basé sur le périmètre (PeCO) qui est aussi un protocole distribué sur les n\oe uds de capteurs dans chaque sous-région. Notre contribution dans ce protocole consiste essentiellement dans la proposition d'un nouveau modèle mathématique de l'optimisation basé sur le périmètre de couverture pour l'ordonnancement de l'activité des capteurs. Un nouveau programme entier du modèle de couverture est résolu par le leader durant la phase de décision pour définir un ensemble de capteurs de couverture pour la phase de surveillance. -Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNET++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. +Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNeT++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. -\textbf{MOTS-CLÉS:} Réseaux sans fil, Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation, Ordonnancement, Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Algorithmes parallèles, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes. +\textbf{MOTS-CLÉS:} Réseaux sans fil, Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation, Ordonnancement, Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes, Simulation des Réseaux, Evaluation de Performance, Les Communications sans Fil Ecologiques et le Réseautage. diff --git a/Thesis.tex b/Thesis.tex index 94c4f1f..8ef577b 100644 --- a/Thesis.tex +++ b/Thesis.tex @@ -32,7 +32,6 @@ \addcontentsline{toc}{chapter}{List of Algorithms} \setlength{\parindent}{0.5cm} - %\addcontentsline{toc}{chapter}{List of Abbreviations} %% Remerciements \include{ACRONYMS} diff --git a/entete.tex b/entete.tex index 7516066..3b17d31 100644 --- a/entete.tex +++ b/entete.tex @@ -74,7 +74,7 @@ \addjury{x1}{y1}{Examiner}{Professor at University of} \addjury{x2}{y2}{Examiner}{Professor at University of} \addjury{x3}{y3}{Examiner}{Professor at University of} -\addjury{x4}{y4}{Examiner}{Professor at University of} +%\addjury{x4}{y4}{Examiner}{Professor at University of} \addjury{Raphaël}{Couturier}{Supervisor}{Professor at University of Franche-Comt\'e} \addjury{Karine}{Deschinkel}{Co-Supervisor}{Assistant Prof. at University of Franche-Comt\'e} \addjury{Michel}{Salomon}{Co-Supervisor}{Assistant Prof. at University of Franche-Comt\'e} @@ -86,19 +86,20 @@ %%-------------------- %% Set the English abstract \thesisabstract[english]{ - +In this dissertation, we highly focus on the area coverage problem, energy-efficiency is also the foremost requirement. We have considered distributed optimization protocols with the ultimate objective of prolonging the network lifetime. The proposed distributed optimization protocols (including algorithms, models, and solving integer programs) should be energy-efficient protocols. To address this problem, this dissertation proposes two-step approaches. Firstly, the sensing field is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of our proposed distributed optimization protocols is distributed and applied on the sensor nodes in each subregion so as to optimize the coverage and the lifetime performances. In this dissertation, three coverage optimization protocols are proposed. These protocols combine two efficient techniques: leader election for each subregion, followed by an optimization-based planning of sensor activity scheduling decisions for each subregion. Extensive simulations are conducted using the discrete event simulator OMNeT++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase the WSN lifetime and provide improved coverage performance. } -\thesiskeywords[english]{ } +\thesiskeywords[english]{ Wireless Networks, Wireless Sensor Networks, Area Coverage, Network Lifetime, Optimization, Scheduling, Distributed Algorithms, Centralized Algorithms, Robustness, Connectivity, Energy-efficiency, Heterogeneous Energy Network, Homogeneous Network, Network Simulation, Performance Evaluation, Wireless Green Communications and Networking.} %%-------------------- %% Set the French abstract \thesisabstract[french]{ +Dans cette thèse, nous nous sommes intéressés au problème de la zone de couverture ainsi qu'à l'efficacité énergétique qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Les protocoles d'optimisation distribués proposés (y compris les algorithmes, les modèles et la résolution des programmes entiers) doivent être efficaces en terme d'énergie. Pour résoudre ce problème, nous avons proposé de nouvelles approches en deux phases. Dans un premier temps, le champ de surveillance est divisé en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Ensuite, l'un de nos protocoles d'optimisation distribués est appliqué sur les n\oe uds de capteurs dans chaque sous-régions afin d'optimiser la couverture et la durée de vie du réseau. Dans cette thèse, nous avons proposé trois protocoles distribués pour l'optimisation de la couverture. Ces protocoles permettent de combiner deux techniques efficaces: une élection de leader pour chaque sous-région, suivie par un processus d'optimisation de l'ordonnancement d'activité de décisions des capteurs pour chaque sous-région. Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNeT++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. } -\thesiskeywords[french]{ } +\thesiskeywords[french]{Réseaux sans fil, Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation, Ordonnancement, Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes, Simulation des Réseaux, Evaluation de Performance, Les Communications sans Fil Ecologiques et le Réseautage. } %%-------------------- -- 2.39.5 From 2357863679fe0c8c15b85903f16b2fc13f553811 Mon Sep 17 00:00:00 2001 From: ali Date: Wed, 6 May 2015 21:46:39 +0200 Subject: [PATCH 10/16] Update by Ali --- Resume.tex | 21 +++++++++++---------- 1 file changed, 11 insertions(+), 10 deletions(-) diff --git a/Resume.tex b/Resume.tex index 5dd0a97..8fccf0d 100644 --- a/Resume.tex +++ b/Resume.tex @@ -4,36 +4,37 @@ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% %% -%% Résumé %% +%% Résumé %%% n\oe ud %% %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -\emph{ \begin{center} \Large Techniques Distribuées d'Optimisation de la Couverture des Réseaux de Capteurs sans Fil pour Améliorer leur Durée de Vie \end{center}} +\emph{ \begin{center} \Large Techniques d'Optimisation Distribuées de la Couverture pour Améliorer la Durée de Vie des Réseaux de Capteurs sans Fil \end{center}} %\emph{ \begin{center} \large By \end{center}} \emph{ \begin{center} \large Ali Kadhum Idrees \\ Université de Franche-Comt\'e, 2015 \end{center}} %\emph{ \begin{center} \large The University of Franche-Comt\'e, 2015 \end{center}} -\emph{ \begin{center} \large Encadrants: Raphaël Couturier, Karine Deschinkel, and Michel Salomon \end{center}} +\emph{ \begin{center} \large Encadrants: Raphaël Couturier, Karine Deschinkel et Michel Salomon \end{center}} -Les réseaux de capteurs sans fil ont suscité beaucoup de travaux de recherche au cours des dernières années en raison de leur large gamme d'applications potentielles. Les caractéristiques des noeuds capteurs imposent des contraints enterme de consommation d'énergie et de capacité de traitement qui rendent caduque les protocoles des réseaux ad-hoc sans fil, avec de nombreux défis à résoudre. Parmi ces défis, on peut noter la préservation de la couverture, le contrôle de la topologie, le routage, la fusion de données, la sécurité, etc. La préservation de la couverture d'une région à surveiller, de manière permanente et efficace, tout en empêchant autant que possible un dysfonctionnement du réseau en raison du déchargement de la batterie de certains n\oe uds, est une des problématique de recherche majeures. +Les réseaux de capteurs sans fil ont suscité beaucoup de travaux de recherche au cours des dernières années en raison de leur large gamme d'applications potentielles. Les caractéristiques des n\oe uds capteurs imposent des contraints enterme de consommation d'énergie et de capacité de traitement qui rendent caduque les protocoles des réseaux ad-hoc sans fil, avec de nombreux défis à résoudre. Parmi ces défis, on peut noter la préservation de la couverture, le contrôle de la topologie, le routage, la fusion de données, la sécurité, etc. La préservation de la couverture d'une région à surveiller, de manière permanente et efficace, tout en empêchant autant que possible un dysfonctionnement du réseau en raison du déchargement de la batterie de certains n\oe uds, est une des problématique de recherche majeures. -Dans cette thèse, nous nous sommes intéressés au problème de la préservation de la couverture, ainsi qu'à l'efficatité qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie opérationnelle du réseau. Les protocoles proposés doivent être efficaces en terme de consommation énergétique induite par les calculs et les communications. Pour résoudre le problème, nous avons proposé des nouvelles approches en deux étapes. Dans un premier temps, la région à surveiller est divisée en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Dans un second temps, un de nos protocoles est exécuté par chacun des noeuds capteurs dans chaque sous-région, afin d'optimiser la couverture et la durée de vie du réseau. Nous proposons trois protocoles distribués qui combinent, chacun, deux techniques efficaces: l'élection d'un noeud leader dans chaque sous-région, suivie par la mise en oeuvre par celui-ci d'un processus de décision via l'optimisation de l'ordonnancement d'activité des noeuds capteurs de sa sous-région. +Dans cette thèse, nous nous sommes intéressés au problème de la préservation de la couverture, ainsi qu'à l'efficatité qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie opérationnelle du réseau. Les protocoles proposés doivent être efficaces en terme de consommation énergétique induite par les calculs et les communications. Pour résoudre le problème, nous avons proposé des nouvelles approches en deux étapes. Dans un premier temps, la région à surveiller est divisée en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Dans un second temps, un de nos protocoles est exécuté par chacun des n\oe uds capteurs dans chaque sous-région, afin d'optimiser la couverture et la durée de vie du réseau. Nous proposons trois protocoles distribués qui combinent, chacun, deux techniques efficaces: l'élection d'un n\oe ud leader dans chaque sous-région, suivie par la mise en oeuvre par celui-ci d'un processus de décision via l'optimisation de l'ordonnancement d'activité des n\oe uds capteurs de sa sous-région. +Le premier protocole proposé est appelé DiLCO, pour Distributed Lifetime Coverage Optimization. Dans ce protocole, la durée de vie est divisée en périodes, avec chaque période qui est composée de 4 phases: échange d'informations entre les n\oe uds d'une sous-région, élection d'un n\oe ud leader, décision et surveillance. Le processus de décision est mis en oeuvre par le n\oe ud leader en résolvant un programme linéaire en nombres entiers qui permet de définir un seul ensemble de n\oe uds de capteurs devant être actifs pour assurer la couverture durant la période courante. +Dans le second protocole, qui est une évolution de DiLCO, nous cherchons à construire simultanément plusieurs ensembles de n\oe uds de capteurs de couverture pour la phase de surveillance. Cette dernière est ainsi diviseé en "rondes" de surveillance, d'où le nom Multiround DiLCO ou MuDiLCO donné à ce protocole. Le processus de décision est toujours effectué par un n\oe ud leader, qui détermine les ensembles de n\oe uds capteurs à activer successivement via la résolution d'un nouveau programme linéaire en nombres entiers. -Le premier protocole proposé est appelé DiLCO, pour Distributed Lifetime Coverage Optimization. Dans ce protocole, la durée de vie est divisée en périodes. Chaque période se compose de quatre phases: échange d'informations, élection de leader, décision et surveillance. Le processus de décision est effectué par le n\oe ud leader, qui résout un programme entier permettant de définir un seul ensemble de n\oe uds de capteurs actifs pour assurer la couverture durant une période. +%Ensuite, nous avons étudié le problème de l'optimisation multi-ronde de la zone de couverture dans un réseau de capteurs sans fil. Nous avons proposé le protocole d'optimisation multi-ronde distribué de la durée de vie de couverture (MuDiLCO) pour étudier la possibilité de fournir plusieurs ensembles de n\oe uds de capteurs de couverture pour la phase de surveillance. Ce protocole travaille également en périodes pendant lesquelles les ensembles de capteurs sont programmés pour rester actifs pour un certain nombre de rondes durant la phase de surveillance, pour assurer la couverture et maximiser la durée de vie du réseau. Le processus de décision est toujours effectué par le n\oe ud leader qui résout un programme entier pour définir un meilleur ensemble de capteurs à être utilisé pendant les rondes de la phase de surveillance. -Ensuite, nous avons étudié le problème de l'optimisation multi-ronde de la zone de couverture dans un réseau de capteurs sans fil. Nous avons proposé le protocole d'optimisation multi-ronde distribué de la durée de vie de couverture (MuDiLCO) pour étudier la possibilité de fournir plusieurs ensembles de n\oe uds de capteurs de couverture pour la phase de surveillance. Ce protocole travaille également en périodes pendant lesquelles les ensembles de capteurs sont programmés pour rester actifs pour un certain nombre de rondes durant la phase de surveillance, pour assurer la couverture et maximiser la durée de vie du réseau. Le processus de décision est toujours effectué par le n\oe ud leader qui résout un programme entier pour définir un meilleur ensemble de capteurs à être utilisé pendant les rondes de la phase de surveillance. - -Enfin, nous avons proposé le protocole d'optimisation de la couverture basé sur le périmètre (PeCO) qui est aussi un protocole distribué sur les n\oe uds de capteurs dans chaque sous-région. Notre contribution dans ce protocole consiste essentiellement dans la proposition d'un nouveau modèle mathématique de l'optimisation basé sur le périmètre de couverture pour l'ordonnancement de l'activité des capteurs. Un nouveau programme entier du modèle de couverture est résolu par le leader durant la phase de décision pour définir un ensemble de capteurs de couverture pour la phase de surveillance. +Enfin, nous avons proposé un protocole d'optimisation de la couverture basé sur le périmètre des n\oe uds de capteurs (PeCO), qui est aussi un protocole distribué sur les n\oe uds de capteurs dans chaque sous-région. Notre contribution dans ce protocole consiste essentiellement dans la proposition d'un nouveau modèle mathématique de l'optimisation basé sur le périmètre de couverture pour l'ordonnancement de l'activité des capteurs. Un nouveau programme entier du modèle de couverture est résolu par le leader durant la phase de décision pour définir un ensemble de capteurs de couverture pour la phase de surveillance. Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNeT++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. -\textbf{MOTS-CLÉS:} Réseaux sans fil, Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation, Ordonnancement, Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes, Simulation des Réseaux, Evaluation de Performance, Les Communications sans Fil Ecologiques et le Réseautage. +\textbf{MOTS-CLÉS:} Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation Distribué, Ordonnancement. +%Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes, Simulation des Réseaux, Evaluation de Performance, Les Communications sans Fil Ecologiques et le Réseautage \ No newline at end of file -- 2.39.5 From b90ddc92fe317cadc93b6130e57ef6368cd53569 Mon Sep 17 00:00:00 2001 From: ali Date: Fri, 8 May 2015 02:19:21 +0200 Subject: [PATCH 11/16] Update by Ali --- CONCLUSION.tex | 2 +- INTRODUCTION.tex | 2 +- entete.tex | 8 ++++---- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/CONCLUSION.tex b/CONCLUSION.tex index 6706da0..fb930c9 100644 --- a/CONCLUSION.tex +++ b/CONCLUSION.tex @@ -7,7 +7,7 @@ \section{Conclusion} -In this dissertation, we have concentrated on on the design of distributed optimization protocols so as to prolong the lifetime of wireless sensor networks. We have addressed the problem of the area coverage and the lifetime optimization in wireless sensor networks. The ultimate goal is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area of interest. +In this dissertation, we have concentrated on the design of distributed optimization protocols so as to prolong the lifetime of wireless sensor networks. We have addressed the problem of the area coverage and the lifetime optimization in wireless sensor networks. The ultimate goal is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area of interest. The first part of the dissertation has presented the scientific background including WSNs, brief survey of related works, and evaluation tools as well as optimization solvers. diff --git a/INTRODUCTION.tex b/INTRODUCTION.tex index 847f097..715cab3 100644 --- a/INTRODUCTION.tex +++ b/INTRODUCTION.tex @@ -43,7 +43,7 @@ The main contributions in this dissertation concentrate on designing distributed \item We devise a framework to schedule nodes to be activated alternatively such that the network lifetime is prolonged while ensuring that a certain level of coverage is preserved. A key idea in our framework is to exploit a spatial-temporal subdivision. On the one hand, the area of interest is divided into several smaller subregions.On the other hand, the timeline is divided into periods of equal length. In each subregion, the sensor nodes will cooperatively choose a leader which will schedule nodes' activities, and this grouping of sensors is similar to typical cluster architecture. -\item We design, called the Distributed Lifetime Coverage Optimization (DILCO) protocol, which maintains the coverage and improves the lifetime in WSNs. DILCO protocol is presented in chapter 4. It is an extension of our approach introduced in \cite{ref159}. In \cite{ref159}, the protocol is deployed over only two subregions. In DILCO protocol, the area of interest is first divided into subregions using a divide-and-conquer algorithm and an activity scheduling for sensor nodes is then planned by the elected leader in each subregion. In fact, the nodes in a subregion can be seen as a cluster where each node sends sensing data to the cluster head or the sink node. Furthermore, the activities in a subregion/cluster can continue even if another cluster stops due to too many node failures. DiLCO protocol considers periods, where a period starts with a discovery phase to exchange information between sensors of the same subregion, in order to choose in a suitable manner a sensor node (the leader) to carry out the coverage strategy. In each subregion, the activation of the sensors for the sensing phase of the current period is obtained by solving an integer program. The resulting activation vector is broadcasted by the leader to every node of its subregion. +\item We design a protocol, called the Distributed Lifetime Coverage Optimization (DILCO) protocol, which maintains the coverage and improves the lifetime in WSNs. DILCO protocol is presented in chapter 4. It is an extension of our approach introduced in \cite{ref159}. In \cite{ref159}, the protocol is deployed over only two subregions. In DILCO protocol, the area of interest is first divided into subregions using a divide-and-conquer algorithm and an activity scheduling for sensor nodes is then planned by the elected leader in each subregion. In fact, the nodes in a subregion can be seen as a cluster where each node sends sensing data to the cluster head or the sink node. Furthermore, the activities in a subregion/cluster can continue even if another cluster stops due to too many node failures. DiLCO protocol considers periods, where a period starts with a discovery phase to exchange information between sensors of the same subregion, in order to choose in a suitable manner a sensor node (the leader) to carry out the coverage strategy. In each subregion, the activation of the sensors for the sensing phase of the current period is obtained by solving an integer program. The resulting activation vector is broadcasted by the leader to every node of its subregion. \item %We extend our work that explained in chapter 4 and present a generalized framework that can be applied to provide the cover sets of all rounds in each period. The MuDiLCO protocol for Multiround Distributed Lifetime Coverage Optimization protocol, presented in chapter 5, is an extension of the approach introduced in chapter 4. In DiLCO protocol, the activity scheduling based optimization is planned for each subregion periodically only for one sensing round. Whilst, we study the possibility of dividing the sensing phase into multiple rounds. In fact, we make a multiround optimization, while it was a single round optimization in our previous contribution. diff --git a/entete.tex b/entete.tex index 3b17d31..68e0256 100644 --- a/entete.tex +++ b/entete.tex @@ -86,20 +86,20 @@ %%-------------------- %% Set the English abstract \thesisabstract[english]{ -In this dissertation, we highly focus on the area coverage problem, energy-efficiency is also the foremost requirement. We have considered distributed optimization protocols with the ultimate objective of prolonging the network lifetime. The proposed distributed optimization protocols (including algorithms, models, and solving integer programs) should be energy-efficient protocols. To address this problem, this dissertation proposes two-step approaches. Firstly, the sensing field is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of our proposed distributed optimization protocols is distributed and applied on the sensor nodes in each subregion so as to optimize the coverage and the lifetime performances. In this dissertation, three coverage optimization protocols are proposed. These protocols combine two efficient techniques: leader election for each subregion, followed by an optimization-based planning of sensor activity scheduling decisions for each subregion. Extensive simulations are conducted using the discrete event simulator OMNeT++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase the WSN lifetime and provide improved coverage performance. +In this dissertation, we focus on the area coverage problem, energy-efficiency is also the foremost requirement. We have considered distributed optimization protocols with the ultimate objective of prolonging the network lifetime. To address this problem, this dissertation proposes two-step approaches. Firstly, the sensing field is divided into smaller subregions using the concept of divide-and-conquer method. Secondly, one of our proposed distributed optimization protocols is distributed and applied on the sensor nodes in each subregion so as to optimize the coverage and the lifetime performances. Three coverage optimization protocols are proposed, They combine two efficient techniques: leader election for each subregion, followed by an optimization-based planning of sensor activity scheduling for each subregion. Extensive simulations are conducted using the discrete event simulator OMNeT++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase the WSN lifetime and provide improved coverage performance. } -\thesiskeywords[english]{ Wireless Networks, Wireless Sensor Networks, Area Coverage, Network Lifetime, Optimization, Scheduling, Distributed Algorithms, Centralized Algorithms, Robustness, Connectivity, Energy-efficiency, Heterogeneous Energy Network, Homogeneous Network, Network Simulation, Performance Evaluation, Wireless Green Communications and Networking.} +\thesiskeywords[english]{ Wireless Sensor Networks, Area Coverage, Network Lifetime, Distributed Optimization, Scheduling.} %%-------------------- %% Set the French abstract \thesisabstract[french]{ -Dans cette thèse, nous nous sommes intéressés au problème de la zone de couverture ainsi qu'à l'efficacité énergétique qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudiés les protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Les protocoles d'optimisation distribués proposés (y compris les algorithmes, les modèles et la résolution des programmes entiers) doivent être efficaces en terme d'énergie. Pour résoudre ce problème, nous avons proposé de nouvelles approches en deux phases. Dans un premier temps, le champ de surveillance est divisé en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Ensuite, l'un de nos protocoles d'optimisation distribués est appliqué sur les n\oe uds de capteurs dans chaque sous-régions afin d'optimiser la couverture et la durée de vie du réseau. Dans cette thèse, nous avons proposé trois protocoles distribués pour l'optimisation de la couverture. Ces protocoles permettent de combiner deux techniques efficaces: une élection de leader pour chaque sous-région, suivie par un processus d'optimisation de l'ordonnancement d'activité de décisions des capteurs pour chaque sous-région. Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNeT++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. +Dans cette thèse, nous nous sommes intéressé au problème de la zone de couverture ainsi qu'à l'efficacité énergétique qui est une exigence essentielle dans un réseau de capteurs sans fil. Nous avons étudié des protocoles d'optimisation distribués avec l'objectif ultime de prolonger la durée de vie du réseau. Pour résoudre le problème, nous avons proposé de nouvelles approches en deux phases. Dans un premier temps, la région à surveiller est divisée en petites sous-régions en utilisant le concept de la méthode diviser pour mieux régner. Ensuite, l'un de nos protocoles d'optimisation distribués est exécuté par chaque n\oe ud capteur dans chaque sous-région, afin d'optimiser la couverture et la durée de vie du réseau. Nous proposons trois protocoles distribués qui combinent, chacun, deux techniques efficaces: l'élection d'un n\oe ud leader dans chaque sous-région, suivie par la mise en oeuvre par celui-ci d'un processus de décision via l'optimisation de l'ordonnancement d'activité des n\oe uds capteurs de sa sous-région. Nous avons effectué plusieurs simulations en utilisant le simulateur à évènements discrets OMNeT++ pour valider l'efficacité de nos protocoles proposés. Nous avons pris en considération les caractéristiques d'un capteur Medusa II pour la consommation d'énergie et le temps de calcul. En comparaison avec deux autres méthodes existantes, nos protocoles ont la capacité d'augmenter la durée de vie du réseau de capteurs et d'améliorer les performances de couverture. } -\thesiskeywords[french]{Réseaux sans fil, Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation, Ordonnancement, Algorithmes distribués, Algorithmes centralisés, Robustesse, Connectivité, Efficacité énergétique, \'Energie des réseaux hétérogènes, Réseaux homogènes, Simulation des Réseaux, Evaluation de Performance, Les Communications sans Fil Ecologiques et le Réseautage. } +\thesiskeywords[french]{Réseaux de capteurs sans fil, Zone de couverture, Durée de vie du réseau, Optimisation distribué, Ordonnancement. } %%-------------------- -- 2.39.5 From 7976cef0966cf4550fcfd57c22cebce39f5fc3c3 Mon Sep 17 00:00:00 2001 From: ali Date: Fri, 8 May 2015 17:26:38 +0200 Subject: [PATCH 12/16] Update by Ali --- Abstruct.tex | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Abstruct.tex b/Abstruct.tex index cfa6033..c8e686d 100644 --- a/Abstruct.tex +++ b/Abstruct.tex @@ -36,7 +36,7 @@ during which sets of sensor nodes are scheduled to remain active for a number of Last but not least, we propose a Perimeter-based Coverage Optimization (PeCO) protocol which is also distributed among sensor nodes in each subregion.The novelty of our approach lies essentially in the formulation of a new mathematical optimization model based on a perimeter coverage level to schedule sensors' activities, whereas we used primary points coverage model in the two previous models. A new integer program coverage model is solved by the leader during the decision phase so as to provide only one cover set of sensors for the sensing phase. -Extensive simulations are conducted using the discrete event simulator OMNET++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase +Extensive simulations are conducted using the discrete event simulator OMNeT++ to validate the efficiency of each of our proposed protocols. We refer to the characteristics of a Medusa II sensor for the energy consumption and the time computation. In comparison with two other existing methods, our protocols are able to increase the WSN lifetime and provide improved coverage performance. -- 2.39.5 From d884c983d0b0fa2e556454780ccf3fa32c6d9b5c Mon Sep 17 00:00:00 2001 From: ali Date: Mon, 11 May 2015 18:32:39 +0200 Subject: [PATCH 13/16] Update by Ali --- CHAPITRE_02.tex | 323 +++++++++++++++++++++++------------------------ INTRODUCTION.tex | 6 +- Thesis.toc | 8 +- bib.bib | 10 ++ 4 files changed, 175 insertions(+), 172 deletions(-) diff --git a/CHAPITRE_02.tex b/CHAPITRE_02.tex index d13ee68..29423d6 100644 --- a/CHAPITRE_02.tex +++ b/CHAPITRE_02.tex @@ -12,9 +12,10 @@ The main objective of deploying a large number of wireless sensor nodes in the target area of interest is to construct a WSN, which is responsible for monitoring the sensing field. The coverage problem represents the principle requirement in these applications. The main question shared by these applications is how can the deployed wireless sensor nodes monitor the physical phenomenon properly. The coverage can be considered as one of the QoS (Quality of Service) parameters, and it is closely related to energy consumption. It represents the sensing task supplied by the wireless sensors in WSNs. -The energy resource limitation of wireless sensor nodes has been considered as a big challenge. So, it is desired to operate the WSN with less energy consumption whilst fulfilling the coverage requirement. The main objective of scattering the wireless sensor nodes over the area of interest is to collect the sensed data of the physical phenomena for processing or reporting, where there are two types of reporting for sensed data in WSNs~\cite{ref138}: event-driven and on-demand. In the latter, the monitoring base station start the reporting operation by transmitting a request to the wireless sensor nodes so as to send their sensed data to the base station; like in inventory tracking application. In the former, the reporting operation is triggered by one or more wireless sensor nodes within the physical phenomena by transmitting their sensed data to the controlling base station; for instance, the forest fire detection application. In hybrid scheme of the two types is more flexible. +The energy resource limitation of wireless sensor nodes has been considered as a big challenge. So, it is desired to operate the WSN with less energy consumption whilst fulfilling the coverage requirement. The main objective of scattering the wireless sensor nodes over the area of interest is to collect the sensed data of the physical phenomena for processing or reporting, where there are two types of reporting for sensed data in WSNs~\cite{ref138}: event-driven and on-demand. In the latter, the monitoring base station starts the reporting operation by transmitting a request to the wireless sensor nodes so as to send their sensed data to the base station; like in inventory tracking application. In the former, the reporting operation is triggered by one or more wireless sensor nodes within the physical phenomena by transmitting their sensed data to the controlling base station; for instance, the forest fire detection application. +%In hybrid scheme of the two types is more flexible. -The ultimate goal of the coverage is to ensure that each point in the sensing field is within the sensing range of at least one sensor node. Some applications require high reliability to perform their tasks, so they need that every point in the sensing field is covered by more than one sensor node. In order to avoid a lack of monitoring in the area of interest, it is necessary that WSNs are deployed with high density so as to exploit the overlapping among the sensor nodes and to prevent malfunction of sensor nodes in severe environments. The overlap can be exploited by choosing the minimum number of sensor nodes to perform the main tasks of the WSN in the sensing field and putting the remaining sensor nodes in very low power sleep mode so as to prolong the network lifetime. This exploitation manner, which is called sensor activity scheduling, aims to set the activity state of each sensor node in the WSN so that the sensing field can be monitored for as long as possible. The required level of coverage should be guaranteed by the activity-based scheduling scheme~\cite{ref139}. Many scheduling algorithms have been described in~\cite{ref58,ref57}. +The ultimate goal of the coverage is to ensure that each point in the sensing field is within the sensing range of at least one sensor node. Some applications require high reliability to perform their tasks, so they need that each point in the sensing field is covered by more than one sensor node. In order to avoid a lack of monitoring in the area of interest, it is necessary that WSNs are deployed with high density so as to exploit the overlapping among the sensor nodes and to prevent malfunction of sensor nodes in severe environments. The overlap can be exploited by choosing the minimum number of sensor nodes to perform the main tasks of the WSN in the sensing field and putting the remaining sensor nodes in very low power sleep mode so as to prolong the network lifetime. This exploitation manner, which is called sensor activity scheduling, aims to set the activity state of each sensor node in the WSN so that the sensing field can be monitored for as long as possible. The required level of coverage should be guaranteed by the activity-based scheduling scheme~\cite{ref139}. Many scheduling algorithms have been described in~\cite{ref58,ref57}. %This dissertation focuses on the problem of covering the area of interest as long as possible. Several proposed approaches to extend the network lifetime whilst maintaining the coverage have been viewed in this chapter. M. Cardei and J. Wu~\cite{ref113} have been surveyed the different coverage formulation models and their assumptions, as well as the solutions provided. In~\cite{ref105}, several coverage problems are presented from different angles, where the models and assumptions, as well as proposed solutions in the literatures, are described. In this dissertation, the main contribution of previous works that deal with the coverage problem have been addressed. We end this chapter by focusing on two algorithms, GAF~\cite{GAF} and DESK~\cite{DESK}, since they have been used for comparison against our coverage protocols. @@ -22,12 +23,8 @@ The ultimate goal of the coverage is to ensure that each point in the sensing fi %\section{Coverage Algorithms} %\label{ch2:sec:02} -\indent This chapter is dedicated to the various approaches proposed in the -literature for the coverage lifetime maximization problem, where the objective -is to optimally schedule sensors' activities in order to extend network lifetime -in WSNs. -In~\cite{ref105}, several coverage problems are presented from different angles, where the models and assumptions, as well as proposed solutions in the literatures, are described. -M. Cardei and J. Wu~\cite{ref113} survey the different coverage formulation models and their assumptions, as well as the solutions provided. They provide a taxonomy for coverage algorithms in WSNs according to several design choices: +\indent This chapter is dedicated to the various approaches proposed in the literature for the coverage lifetime maximization problem, where the objective +is to optimally schedule sensors' activities in order to extend network lifetime in WSNs. In~\cite{ref105}, several coverage problems are presented from different points of view, where the models and assumptions, as well as proposed solutions in the literatures, are described. M. Cardei and J. Wu~\cite{ref113} survey the different coverage formulation models and their assumptions, as well as the solutions provided. They provide a taxonomy for coverage algorithms in WSNs according to several design choices: \begin{enumerate} [(i)] \item Sensors scheduling algorithm implementation, i.e. centralized or distributed/localized algorithms. @@ -41,13 +38,15 @@ M. Cardei and J. Wu~\cite{ref113} survey the different coverage formulation mode From our point of view, the choice of non-disjoint or disjoint cover sets (sensors participate or not in many cover sets), coverage type (area, target, or barrier), coverage ratio, coverage degree (how many sensors are required to cover a target or an area) can be added to the above list. -Once sensor nodes are deployed, a coverage algorithm is run to schedule the sensor nodes into cover sets so as to maintain sufficient coverage in the area of interest and extend the network lifetime. The WSN applications require either complete or partial area coverage, while for target coverage, all the target should be covered. This chapter concentrates only on area coverage and target coverage problems because it is possible to transform the area coverage problem to target (or point) coverage problem and vice versa. We have excluded the barrier coverage problem from this discussion because it is outside the scope of this dissertation. -This dissertation focuses mainly on the area coverage problem, where the ultimate goal is to choose the minimum number of sensor nodes to cover the whole sensing field. +Once sensor nodes are deployed, a coverage algorithm is run to schedule the sensor nodes into cover sets so as to maintain sufficient coverage in the area of interest and extend the network lifetime. +%The WSN applications require either complete or partial area coverage, while for target coverage, all the target should be covered. +This chapter concentrates only on area coverage and target coverage problems because it is possible to transform the area coverage problem to target (or point) coverage problem and vice versa. We have excluded the barrier coverage problem from this discussion because it is outside the scope of this dissertation. +This dissertation mainly focuses on the area coverage problem, where the ultimate goal is to choose the minimum number of sensor nodes to cover the whole sensing field. %We have focused mainly on the area coverage problem. Therefore, we represent the sensing area of each sensor node in the sensing field as a set of primary points and then achieving full area coverage by covering all the points in the sensing field. The ultimate goal of the area coverage problem is to choose the minimum number of sensor nodes to cover the whole sensing region and prolonging the lifetime of the WSN. -Many centralized and distributed coverage algorithms for activity scheduling have been proposed in the literature, based on different assumptions and objectives. In centralized algorithms, a central controller makes all decisions and distributes the results to sensor nodes. The centralized algorithms have the advantage of requiring very low processing power from the sensor nodes which have usually limited processing capabilities. On the contrary, the exchange of packets in large WSNs may consume a considerable amount of energy in a centralized approach compared to a distributed one. Moreover, centralized approaches usually suffer from the scalability and reliability problems, making them less competitive as the network size increases. +Many centralized and distributed coverage algorithms for activity scheduling have been proposed in the literature, based on different assumptions and objectives. In centralized algorithms, a central controller makes all decisions and distributes the results to sensor nodes. The centralized algorithms have the advantage of requiring very low processing power from the sensor nodes which have usually limited processing capabilities. On the contrary, the exchange of packets in large WSNs may consume a considerable amount of energy in a centralized approach compared to a distributed one. Moreover, centralized approaches usually suffer from the scalability and reliability problems, making them less competitive than the network size increases. -In a distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and only informations from neighboring nodes are used for the activity decision. Compared to centralized algorithms, distributed algorithms reduce the energy consumption required for radio communication and detection accuracy whilst the energy consumption for computation is increased. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give an optimal (or near-optimal) solution based only on local informations. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. Table~\ref{Table0:ch2} shows a comparison between centralized coverage algorithms and distributed coverage algorithms. +In distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and only informations from neighboring nodes are used for the activity decision. Compared to centralized algorithms, distributed algorithms reduce the energy consumption required for radio communication and detection accuracy whilst the energy consumption for computation is increased. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give an optimal (or near-optimal) solution based only on local informations. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. Table~\ref{Table0:ch2} shows a comparison between centralized coverage algorithms and distributed coverage algorithms. \begin{table}[h!] \caption{Centralized Coverage Algorithms vs Distributed Coverage Algorithms} @@ -81,41 +80,147 @@ In this dissertation, the sensing field is divided into smaller subregions using Several algorithms to maintain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. Table \ref{x11} summarizes the main characteristics of some coverage approaches in previous literatures. -In this table \ref{x11}, the "SET K-COVER" characteristic refers to the maximum number of disjoint or non-disjoint sets of sensors such that each set cover can assure the coverage for the whole region. -The K-COVER algorithm provides a solution with K cover sets in each execution. The k-coverage characteristic refers to the fact that every point inside the monitored area is always covered by at least k active sensors. +In this table, the "SET K-COVER" characteristic refers to the maximum number of disjoint or non-disjoint sets of sensors such that each set cover can ensure the coverage for the whole region. The K-COVER algorithm provides a solution with K cover sets in each execution. The k-coverage characteristic refers to the fact that every point inside the monitored area is always covered by at least k active sensors. -\section{Centralized Algorithms} -\label{ch2:sec:02} -The major idea of most centralized algorithms is to divide/organize the sensors into a suitable number of cover sets, where each set completely covers an interest region and to activate these cover sets successively. The centralized algorithms always provide optimal or near-optimal solution since the algorithm has a global view of the whole network. Energy-efficient centralized approaches differ according to several criteria \cite{ref113}, such as the coverage objective (target coverage or area coverage), the node deployment method (random or deterministic), and the heterogeneity of sensor nodes (common sensing range, common battery lifetime). +\begin{table}[h!] -The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers such that the number of covers that include an area, summed over all areas, is maximized. -Their work is built upon previous work in~\cite{ref116} and the generated cover sets do not provide complete coverage of the monitoring zone. +\begin{flushleft} +\centering +\caption{Main characteristics of some coverage approaches in literature.} +\label{x11} + \begin{tabular}{@{} cl*{13}c @{}} + & & \\ + & & \multicolumn{10}{c}{Characteristics} \\[2ex] + \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds or Periods} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ + \cmidrule[1pt]{2-14} -The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$, where -$n$ is the number of sensors and $m$ the number of targets. Compared to algorithm's results of Slijepcevic and Potkonjak \cite{ref116}, their heuristic produces more cover sets with a slight growth rate in execution time. L. Liu et al.~\cite{ref150} formulate the maximum disjoint sets for maintaining target coverage and connectivity problem in WSN. They propose a graph theoretical framework to study the joint problem of connectivity and coverage in a WSN with random deployment of nodes with no restrictions on the sensing and communication ranges of nodes. They propose heuristic algorithms to find the suitable number of nodes to guarantee connectivity and coverage while maximizing network lifetime. -%This work did not take into account the sensor node failure, which is an unpredictable event because the two solutions are full centralized algorithms. -Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem to a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find full coverage sets with virtual radii and then transform the coverage sets to a partial coverage sets by adjusting sensing radii. This framework has four strategies, two of them are designed for network, where the sensors have fixed sensing range and the other two are for network, where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{ref229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets, each capable of monitoring all the targets of the region of interest. %Those covers sets are scheduled periodically. -Their algorithm is able to construct the different cover sets in parallel. The results show that their algorithm achieves near-optimal solutions compared to the optimal ones obtained by the resolution of an integer programming problem. -%exact method. +& \tiny Z. Abrams et al. (2004)~\cite{ref114} & \OK &\OK & \OK & & & &\OK & \OK & & \OK & & &\\ + +& \tiny M. Cardei and D. Du (2005)~\cite{ref115} & & \OK & & \OK & & & \OK & \OK & & \OK & & &\\ + +& \tiny S. Slijepcevic and M. Potkonjak (2001)~\cite{ref116} & & \OK & \OK & & & & \OK & \OK & & \OK & & &\\ + +& \tiny Manjun and A. K. Pujari (2011)~\cite{ref117} & & \OK & & \OK & & & \OK & & \OK & & & &\\ + +& \tiny M. Yang and J. Liu (2014)~\cite{ref118} & & \OK & \OK & & & & \OK & & \OK & & & & \\ + +& \tiny S. Wang et al. (2010)~\cite{ref144} & & \OK & \OK & & & & \OK & & \OK & & \OK & & \\ + +& \tiny C. Lin et al. (2010)~\cite{ref147} & & \OK & \OK & & & & \OK & & \OK & & & & \\ + +& \tiny S. A. R. Zaidi et al. (2009)~\cite{ref148} & & \OK & \OK & & & & \OK & & \OK & & & & \\ + +& \tiny Y. Li et al. (2011)~\cite{ref142} & & \OK & \OK & & & \OK & \OK & \OK & & \OK & & \OK &\\ + +& \tiny H. M. Ammari and S. K. Das (2012)~\cite{ref152} & \OK & \OK & \OK & & \OK & & \OK & & \OK & & \OK & &\\ + +& \tiny L. Liu et al. (2010)~\cite{ref150} & & \OK & & \OK & & \OK & & \OK & & \OK & & &\\ + +& \tiny H. Cheng et al. (2014)~\cite{ref119} & & \OK & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny M. Rebai et al. (2014)~\cite{ref141} & & \OK & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny L. Aslanyan et al. (2013)~\cite{ref151} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ + +& \tiny X. Liu et al. (2014)~\cite{ref143} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ + +& \tiny F. Castano et al. (2013)~\cite{ref120} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ + +& \tiny A. Rossi et al. (2012)~\cite{ref121} & & \OK & & \OK & & \OK & \OK & & \OK & \OK & & \OK &\\ + +& \tiny K. Deschinkel et al. (2012)~\cite{ref122} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ + + + +& \tiny A. Gallais et al. (2008)~\cite{ref123} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & \OK &\\ + +& \tiny D. Tian and N. D. Georganas (2002)~\cite{ref124} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny F. Ye et al. (2003)~\cite{ref125} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny H. Zhang and J. C. Hou (2005)~\cite{ref126} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny W. B. Heinzelman et al. (2002)~\cite{ref109} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny T. Yardibi and E. Karasan (2010)~\cite{ref127} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ +& \tiny S. K. Prasad and A. Dhawan (2007)~\cite{ref128} & \OK & & & \OK & & & \OK & & \OK & & \OK & &\\ -In the case of non-disjoint algorithms~\cite{ref117}, sensors may participate in more than one cover set. In some cases, this may prolong the lifetime of the network in comparison to the disjoint cover set algorithms, but designing algorithms for non-disjoint cover sets generally induces a higher order of complexity. Moreover, in case of a sensor's failure, non-disjoint scheduling policies are less resilient and reliable because a sensor may be involved in more than one cover sets. For instance, Cardei et al.~\cite{ref167} -present a Linear Programming (LP) solution and a greedy approach to -extend the sensor network lifetime by organizing the sensors into a -maximal number of non-disjoint cover sets. Simulation results show -that by allowing sensors to participate in multiple sets, the network -lifetime increases compared with related work~\cite{ref115}. The authors in~\cite{ref148}, address the problem of minimum cost area coverage in which full coverage is performed by using the minimum number of sensors for an arbitrary geometric shape region. A geometric solution to the minimum cost coverage problem under a deterministic deployment is proposed. The probabilistic coverage solution which provides a relationship between the probability of coverage and the number of randomly deployed sensors in an arbitrarily-shaped region is suggested. -%The authors explained that with a random deployment about seven times more nodes are required to supply full coverage compared to deterministic deployment. -The work in~\cite{ref144} address the area coverage problem by proposing a Geometrically based Activity Scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explain that GAS is capable to monitor the target area by using the fewest number of sensors and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called Node Coverage Grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other and the data collected by those in the same group are supposed to be similar. They prove that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. -For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be a premature assumption that sensors near to each other sense similar data. The problem of k-coverage over the area of interest in WSNs is addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They check the connection of the graph via laplacian of the adjacency graph of active sensors in each round. %The generation of coverage bitmap by using Minkowski technique, the network is able to providing the desired ratio of coverage. -They define the connected coverage problem as an optimization problem and a centralized genetic algorithm is used to find the solution. +& \tiny S. Misra et al. (2011)~\cite{ref97} & \OK & & \OK & & & & \OK & & \OK & & & &\\ -Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSNs \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of Maximum-network Lifetime Problem (MLP). CG decomposes the problem into a Restricted Master Problem (RMP) and a Pricing Subproblem (PS). The former maximizes lifetime using an incomplete set of columns and the latter is used to identify new profitable columns. A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation boosted by a Greedy Randomized Adaptive Search Procedure (GRASP) and a Variable Neighborhood Search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. +& \tiny P. Berman et al. (2005)~\cite{ref130} & \OK & \OK & \OK & & & & \OK & & \OK & \OK & &\\ -More recently, the authors in~\cite{ref118} consider an area coverage optimization algorithm based on linear programming approach to select the minimum number of working sensor nodes, in order to preserve a maximum coverage and to extend the lifetime of the network. The experimental results show that linear programming can provide a fewest number of active nodes and maximize the network lifetime coverage. M. Rebai et al.~\cite{ref141}, formulate the problem of full grid area coverage problem using two integer linear programming models: the first, is in model that takes into account only the overall coverage constraint; the second, both the connectivity and the full grid coverage constraints are taken into consideration. This work does not consider the energy constraint. H. Cheng et al.~\cite{ref119} define a heuristic area coverage algorithm called Cover Sets Balance (CSB), which chooses a set of active nodes using the tuple (data coverage range, residual energy). Then, they introduce a new Correlated Node Set Computing (CNSC) algorithm to find the correlated node set for a given node. After that, they propose a High Residual Energy First (HREF) node selection algorithm to minimize the number of active nodes. X. Liu et al.~\cite{ref143} explain that in some applications of WSNs such as Structural Health Monitoring (SHM) and volcano monitoring, the traditional coverage model which is a geographic area defined for individual sensors is not always valid. For this reason, they define a generalized area coverage model, which is not required to have the coverage area of individual nodes, but only based on a function deciding whether a set of sensor nodes is capable of satisfy the requested monitoring task for a certain area. They propose two approaches for dividing the deployed nodes into suitable cover sets. +& \tiny J. Lu and T. Suda (2003)~\cite{ref131} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + + + +& \tiny J. Cho et al. (2007)~\cite{ref145} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny V. T. Quang and T. Miyoshi (2008)~\cite{ref146} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ + +%\rot{\rlap{Some Proposed Coverage Protocols in previous literatures}} + +& \tiny D. Dong et al. (2012)~\cite{ref149} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny B. Wang et al. (2012)~\cite{ref134} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny Z. Liu et al. (2012)~\cite{ref135} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny L. Zhang et al. (2013)~\cite{ref136} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & &\\ + +& \tiny S. He et al. (2012)~\cite{ref137} & \OK & \OK & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny Y. Xu et al. (2001)~\cite{GAF} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny C. Vu et al. (2006)~\cite{DESK} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ + +& \tiny X. Deng et al. (2012)~\cite{ref160} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny X. Deng et al. (2005)~\cite{ref133} & \OK & & \OK & & \OK & & \OK & & \OK & & & &\\ + +&\textbf{\textcolor{red}{ \tiny DiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ + +&\textbf{\textcolor{red}{ \tiny MuDiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} &\textbf{\textcolor{red}{\OK}} & & \\ + +&\textbf{\textcolor{red}{ \tiny PeCO Protocol (2015)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ + + \cmidrule[1pt]{2-14} + \end{tabular} + \end{flushleft} + + + +\end{table} + + + + +\section{Centralized Algorithms} +\label{ch2:sec:02} +The major idea of most centralized algorithms is to divide/organize the sensors into a suitable number of cover sets and to activate these cover sets successively. The centralized algorithms always provide optimal or near-optimal solution since the algorithm has a global view of the whole network. Energy-efficient centralized approaches differ according to several criteria \cite{ref113}, such as the coverage objective (target coverage or area coverage), the node deployment method (random or deterministic), and the heterogeneity of sensor nodes (common sensing range, common battery lifetime). + +The first algorithms proposed in the literature consider that the cover sets are disjoint: a sensor node appears in exactly one of the generated cover sets~\cite{ref114,ref115,ref116,ref227}. For instance, Slijepcevic and Potkonjak \cite{ref116} propose an algorithm, which allocates sensor nodes in mutually independent sets to monitor an area divided into several fields. Their algorithm builds a cover set by including in priority the sensor nodes which cover critical fields, that is to say, fields that are covered by the smallest number of sensors. The time complexity of their heuristic is $O(n^2)$ where $n$ is the number of sensors. +%%%M. Cardei et al.~\cite{ref227}, suggest a graph coloring technique to achieve energy savings by organizing the sensor nodes into a maximum number of disjoint dominating sets, which are activated successively. They have defined the maximum disjoint dominating sets problem and they have produced a heuristic that computes the disjoint cover sets so as to monitor the area of interest. The dominating sets do not guarantee the coverage of the whole region of interest. Abrams et al.~\cite{ref114} design three approximation algorithms for a variation of the set k-cover problem, where the objective is to partition the sensors into covers so that the number of covers that include an area, summed over all areas, is maximized. Their work is built upon previous work in~\cite{ref116} and the generated cover sets do not provide complete coverage of the monitoring zone. +%%%The authors in~\cite{ref115} propose a heuristic to compute the Disjoint Set Covers (DSC). In order to compute the maximum number of covers, they first transform DSC into a maximum-flow problem, which is then formulated as a Mixed Integer Programming problem (MIP). Based on the solution of the MIP, they design a heuristic to compute the final number of covers. The results show a slight performance improvement in terms of the number of produced DSC in comparison to~\cite{ref116}, but it incurs higher execution time due to the complexity of the mixed integer programming solving. Zorbas et al. \cite{ref228} present B\{GOP\}, a centralized target coverage algorithm introducing sensor candidate categorization depending on their coverage status and the notion of critical target to call targets that are associated with a small number of sensors. The total running time of their heuristic is $0(m n^2)$, where $n$ is the number of sensors and $m$ the number of targets. Compared to algorithm's results of Slijepcevic and Potkonjak \cite{ref116}, their heuristic produces more cover sets with a slight growth rate in execution time. +L. Liu et al.~\cite{ref150} formulate the maximum disjoint sets for maintaining target coverage and connectivity problem in WSN. They propose a graph theoretical framework to study the joint problem of connectivity and coverage in a WSN with random deployment of nodes with no restrictions on the sensing and communication ranges of nodes. They propose heuristic algorithms to find the suitable number of nodes to guarantee connectivity and coverage while maximizing network lifetime. +Y. Li et al.~\cite{ref142} present a framework with heuristic strategies to solve the area coverage problem. The framework converts any complete coverage problem into a partial coverage one with any coverage ratio. They execute a complete coverage algorithm to find full coverage sets with virtual radii and then transform the coverage sets to a partial coverage sets by adjusting sensing radii. This framework has four strategies, two of them are designed for network, where the sensors have fixed sensing range and the other two are for network, where the sensors have adjustable sensing range. The properties of the algorithms can be maintained by this framework and the transformation process has a low execution time. The simulation results validate the efficiency of the four proposed strategies. More recently, Deschinkel and Hakem \cite{ref229} introduce a near-optimal heuristic algorithm for solving the target coverage problem in WSN. The sensor nodes are organized into disjoint cover sets, each one is capable of monitoring all the targets of the region of interest. Their algorithm is able to construct the different cover sets in parallel. The results show that their algorithm achieves near-optimal solutions compared to the optimal ones obtained by the resolution of an integer programming problem. +%exact method. + + +In the case of non-disjoint algorithms~\cite{ref117,ref167,ref144,ref147,ref118}, sensors may participate in more than one cover set. In some cases, this may prolong the lifetime of the network in comparison to the disjoint cover set algorithms, but designing algorithms for non-disjoint cover sets generally induces a higher order of complexity. Moreover, in case of a sensor's failure, non-disjoint scheduling policies are less resilient and reliable because a sensor may be involved in more than one cover sets. For instance, +%%%Cardei et al.~\cite{ref167} present a Linear Programming (LP) solution and a greedy approach to extend the sensor network lifetime by organizing the sensors into a maximal number of non-disjoint cover sets. Simulation results show that by allowing sensors to participate in multiple sets, the network lifetime increases compared with related work~\cite{ref115}. +The authors in~\cite{ref148}, address the problem of minimum cost area coverage in which full coverage is performed by using the minimum number of sensors for an arbitrary geometric shape region. A geometric solution to the minimum cost coverage problem under a deterministic deployment is proposed. The probabilistic coverage solution which provides a relationship between the probability of coverage and the number of randomly deployed sensors in an arbitrarily-shaped region is suggested. +%%%The work in~\cite{ref144} addresses the area coverage problem by proposing a Geometrically based Activity Scheduling scheme, named GAS, to fully cover the area of interest in WSNs. The authors deal with a small area, called target area coverage, which can be monitored by a single sensor instead of area coverage, which focuses on a large area that should be monitored by many sensors cooperatively. They explain that GAS is capable to monitor the target area by using the fewest number of sensors and it can produce as many cover sets as possible. A novel area coverage method to divide the sensors called Node Coverage Grouping (NCG) is suggested~\cite{ref147}. The sensors in the connectivity group are within sensing range of each other and the data collected by those in the same group are supposed to be similar. They prove that dividing N sensors via NCG into connectivity groups is an NP-hard problem. So, a heuristic algorithm of NCG with time complexity of $O(n^3)$ is proposed. For some applications, such as monitoring an ecosystem with extremely diversified environment, it might be a premature assumption that sensors near to each other sense similar data. +The problem of k-coverage over the area of interest in WSNs is addressed in~\cite{ref152}. It is mathematically formulated and the spatial sensor density for full k-coverage is determined. The relation between the communication range and the sensing range is constructed by this work to retain the k-coverage and connectivity in WSN. After that, four configuration protocols are proposed for treating the k-coverage in WSNs. Simulation results show that their protocols outperform an existing distributed k-coverage configuration protocol. The work presented in~\cite{ref151} solves the area coverage and connectivity problem in sensor networks in an integrated way. The network lifetime is divided into a fixed number of rounds. A coverage bitmap of sensors of the domain is generated in each round and based on this bitmap, it is decided which sensors stay active or go to sleep. They check the connection of the graph via laplacian of the adjacency graph of active sensors in each round. They define the connected coverage problem as an optimization problem and a centralized genetic algorithm is used to find the solution. Recent studies show an increasing interest in the use of exact schemes to solve optimization problems in WSNs \cite{ref230,ref231,ref121,ref122,ref120}. Column Generation (CG) has been widely used to address different versions of Maximum-network Lifetime Problem (MLP). CG decomposes the problem into a Restricted Master Problem (RMP) and a Pricing Subproblem (PS). The former maximizes lifetime using an incomplete set of columns and the latter is used to identify new profitable columns. +%%%A. Rossi et al.~\cite{ref121} introduce an efficient implementation of a genetic algorithm based on CG to extend the lifetime and maximize target coverage in wireless sensor networks under bandwidth constraints. The authors show that the use of metaheuristic methods to solve PS in the context of CG allows to obtain optimal solutions quite fast and to produce high-quality solutions when the algorithm is stopped before returning an optimal solution. More recently, + F. Castano et al. \cite{ref120} address the maximum network lifetime and the target coverage problem in WSNs with connectivity and coverage constraints. They consider two cases to schedule the activity of a set of sensor nodes, keeping them connected while network lifetime is maximized. First, the full coverage of the targets is required, second only a fraction of the targets has to be monitored at any instant of time. They propose an exact approach based on column generation boosted by a Greedy Randomized Adaptive Search Procedure (GRASP) and a Variable Neighborhood Search (VNS) to address both of these problems. Finally, a multiphase framework combining these two approaches is constructed sequentially using these two heuristics at each iteration of the column generation algorithm. The results show that combining the two heuristic methods enhances the results significantly. + +More recently, +%%%the authors in~\cite{ref118} consider an area coverage optimization algorithm based on linear programming approach to select the minimum number of working sensor nodes, in order to preserve a maximum coverage and to extend the lifetime of the network. The experimental results show that linear programming can provide a fewest number of active nodes and maximize the network lifetime coverage. +M. Rebai et al.~\cite{ref141}, formulate the problem of full grid area coverage problem using two integer linear programming models: the first, is in model that takes into account only the overall coverage constraint; the second, both the connectivity and the full grid coverage constraints are taken into consideration. This work does not consider the energy constraint. H. Cheng et al.~\cite{ref119} define a heuristic area coverage algorithm called Cover Sets Balance (CSB), which chooses a set of active nodes using the tuple (data coverage range, residual energy). Then, they introduce a new Correlated Node Set Computing (CNSC) algorithm to find the correlated node set for a given node. After that, they propose a High Residual Energy First (HREF) node selection algorithm to minimize the number of active nodes. X. Liu et al.~\cite{ref143} explain that in some applications of WSNs such as Structural Health Monitoring (SHM) and volcano monitoring, the traditional coverage model which is a geographic area defined for individual sensors is not always valid. For this reason, they define a generalized area coverage model, which is not required to have the coverage area of individual nodes, but only based on a function deciding whether a set of sensor nodes is capable of satisfy the requested monitoring task for a certain area. They propose two approaches for dividing the deployed nodes into suitable cover sets. @@ -128,25 +233,25 @@ More recently, the authors in~\cite{ref118} consider an area coverage optimizati %In distributed and localized coverage algorithms, the required computation to schedule the activity of sensor nodes will be done by the cooperation among neighboring nodes. These algorithms may require more computation power for the processing by the cooperating sensor nodes, but they are more scalable for large WSNs. -Many distributed algorithms have been developed to perform the scheduling so as to preserve coverage, see for example \cite{ref123,ref124,ref125,ref126,ref109,ref127,ref128,ref97}. Localized and distributed algorithms generally result in non-disjoint set covers. +Many distributed algorithms have been developed to perform the scheduling to preserve coverage, see for example \cite{ref123,ref124,ref125,ref126,ref109,ref127,ref128,ref97}. Localized and distributed algorithms generally result in non-disjoint set covers. -X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision problem, whose goal is to determine whether every point in the area of interest is monitored by at least k sensors. The authors prove that if the perimeters of sensors are sufficiently covered it will be the case for the whole area. They provide an algorithm in $O(nd~log~d)$ time to compute the perimeter coverage of each sensor, where $d$ denotes the maximum number of sensors that are neighbors of a sensor and $n$ is the total number of sensors in the network. +X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision problem, whose goal is to determine if all points in the area of interest are monitored by at least k sensors. The authors prove that if the perimeters of sensors are sufficiently covered it will be the case for the whole area. They provide an algorithm in $O(nd~log~d)$ time to compute the perimeter coverage of each sensor, where $d$ denotes the maximum number of sensors that are neighbors of a sensor and $n$ is the total number of sensors in the network. %Their solutions can be translated to distributed protocols to solve the coverage problem. Distributed algorithms typically operate in rounds for a predetermined duration. At the beginning of each round, a sensor exchanges information with its neighbors and makes a decision to either remain turned on or to go to sleep for the round. This decision is basically made on simple greedy criteria like the largest uncovered area \cite{ref130} or maximum uncovered targets \cite{ref131}. Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increases network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfilling the needed sensing coverage. The authors in~\cite{ref146} define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode. In addition, a smaller number of active sensors is chosen so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. -A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is able to build a sparse coverage set in distributed way by means of only connectivity information. This work considers only that the communication range of the sensor is two times smaller than the sensing one. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160} design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disc of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing +A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is able to build a sparse coverage set in distributed way by means of only connectivity information. This work only considers that the communication range of the sensor is two times smaller than the sensing one. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160} design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disc of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing field completely. Simulations results show that this approach can prolong the lifetime of the network compared with other works. The works presented in~\cite{ref134,ref135,ref136} focus on coverage-aware, distributed energy-efficient, and distributed clustering methods respectively, which aim at extending the network lifetime, while the coverage is ensured. -In this dissertation, we focus in more details on two distributed coverage algorithms: GAF and DESK, because we compared our proposed coverage optimization protocols with them during performance evaluation. +In this dissertation, we focus in more details on two distributed coverage algorithms: GAF and DESK, because we compared our proposed coverage optimization protocols with them during performance evaluation. GAF algorithm is chose for comparison as competitor because it is famous, fast, and easy to implement, as well as many authors referred to it in many publications. In addition, DESK algorithm is also selected as competitor in the comparison because it is a full distributed coverage approach. \subsection{Geographical Adaptive Fidelity (GAF)} \label{ch2:sec:03:1} -GAF is developed by Xu et al. \cite{GAF}, it uses geographic location information to divide the area of interest into a fixed square grids. Within each fixed square grid, it keeps only one node staying awake to take the responsibility of sensing and communication. Each sensor node uses its GPS to associate itself with a point in the grid.Figure~\ref{gaf1} gives an example of fixed square grid in GAF. +GAF is developed by Xu et al. \cite{GAF}, it uses geographic location information to divide the area of interest into a fixed square grids. Within each fixed square grid, it keeps only one node staying awake to take the responsibility of sensing and communication. Each sensor node uses its GPS to associate itself with a point in the grid. Figure~\ref{gaf1} gives an example of fixed square grid in GAF. \begin{figure}[h!] \centering @@ -190,12 +295,12 @@ one sensor node (based on the remaining energy of sensor nodes inside the fixed \label{ch2:sec:03:2} % The authors in~\cite{DESK} design a novel distributed heuristic, called Distributed Energy-efficient Scheduling for K-coverage (DESK), which -DESK is a novel distributed heuristic to ensure that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied~\cite{DESK}. This heuristic works in rounds, requires only one-hop neighbor information, and each sensor decides its status (Active or Sleep) based on the perimeter coverage model from~\cite{ref133}. +DESK is a novel distributed heuristic to ensure that the energy consumption among the sensors is balanced and the lifetime maximized while the coverage requirement is satisfied~\cite{DESK}. This heuristic works in rounds, it requires only one-hop neighbor information, and each sensor decides its status (Active or Sleep) based on the perimeter coverage model from~\cite{ref133}. %DESK is based on the result from \cite{ref133}. -In DESK \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are K-covered. The coverage level of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst figure~\ref{figp}~(b) shows the angles corresponding with those arcs in the range [0,2$ \pi $]. According to figure~\ref{figp}~(a) and (b), the coverage level of sensor $s_i$ can be calculated as follows. +In DESK \cite{ref133}, the whole area is K-covered if and only if the perimeters of all sensors are K-covered. The coverage level of a sensor $s_i$ is determined by calculating the angle corresponding to the arc that each of its neighbors covers its perimeter. Figure~\ref{figp}~(a) illuminates such arcs whilst Figure~\ref{figp}~(b) shows the angles corresponding with those arcs in the range [0,2$ \pi $]. According to figure~\ref{figp}~(a) and (b), the coverage level of sensor $s_i$ can be calculated as follows. %via traversing the range from 0 to 2$ \pi $. -For each sensor $s_j$ such that $d(s_i,s_j)$ $<$ $2R_s$, calculate the angle of $s_i$'s arc, denoted by [$\alpha_{j,L}$, $\alpha_{j,R}$], which is perimeter covered by $s_j$, where $\alpha= arccos(d(s_i, s_j)/2R_s)$ and $d(s_i,s_j)$ is the Euclidean distance between $s_i$ and $s_j$. After that, locate the points $\alpha_{j,L}$ and $\alpha_{j,R}$ of each neighboring sensor $s_j$ of $s_i$ on the line segment $[0, 2\pi]$. These points are sorted in ascending order into a list L. Traverse the line segment from 0 to $2\pi$ by visiting each element in the sorted list L from the left to the right and determine the perimeter coverage of $s_i$. Whenever an element $\alpha_{j,L}$ is traversed, the level of perimeter coverage should be increased by one. Whenever an element $\alpha_{j,R}$ is traversed, the level of perimeter coverage should be decreased by one. +For each sensor $s_j$ such that $d(s_i,s_j)$ $<$ $2R_s$, we calculate the angle of $s_i$'s arc, denoted by [$\alpha_{j,L}$, $\alpha_{j,R}$], which is perimeter covered by $s_j$, where $\alpha= arccos(d(s_i, s_j)/2R_s)$ and $d(s_i,s_j)$ is the Euclidean distance between $s_i$ and $s_j$. After that, we locate the points $\alpha_{j,L}$ and $\alpha_{j,R}$ of each neighboring sensor $s_j$ of $s_i$ on the line segment $[0, 2\pi]$. These points are sorted in ascending order into a list L. We traverse the line segment from 0 to $2\pi$ by visiting each element in the sorted list L from the left to the right and determine the perimeter coverage of $s_i$. Whenever an element $\alpha_{j,L}$ is traversed, the level of perimeter coverage should be increased by one. Whenever an element $\alpha_{j,R}$ is traversed, the level of perimeter coverage should be decreased by one. \begin{figure}[h!] @@ -217,7 +322,7 @@ For each sensor $s_j$ such that $d(s_i,s_j)$ $<$ $2R_s$, calculate the angle of \label{desk} \end{figure} -Figure~\ref{desk} shows the DESK network time line. DESK works into rounds fashion. The network lifetime is divided into R rounds. Each round consists of two phases: decision phase and sensing phase. The length of round is dRound that means each sensor node executes this algorithm every dRound unit of time. The decision should be taken within W unit of time, where $W<< dRound$ as shown in figure~\ref{desk}. All the sensor nodes should be temporarily awakened in the decision phase so as to decide their status. Every sensor node $s_i$ decides its status to be active or sleep after $w_i$ of waiting time. The waiting time $w_i$ of node $s_i$ is dynamic and can be changed at any time based on the status of its neighbors, the remaining energy $e_i$ of $s_i$, and its contribution $c_i$ in the coverage level of the network, where $c_i$ is defined as the number of neighbors which need $s_i$ to be active. The waiting time is defined as follows: +Figure~\ref{desk} shows the DESK network time line. DESK works into rounds fashion. The network lifetime is divided into R rounds. Each round consists of two phases: decision phase and sensing phase. The length of round is dRound that means that each sensor node executes this algorithm every dRound unit of time. The decision should be taken within W unit of time, where $W<< dRound$ as shown in Figure~\ref{desk}. All the sensor nodes should be temporarily awakened in the decision phase so as to decide their status. Every sensor node $s_i$ decides its status to be active or sleep after $w_i$ of waiting time. The waiting time $w_i$ of node $s_i$ is dynamic. It can be changed at any time based on the status of its neighbors, the remaining energy $e_i$ of $s_i$, and its contribution $c_i$ in the coverage level of the network. The contribution $c_i$ is defined as the number of neighbors which need $s_i$ to be active. The waiting time is defined as follows: \begin{equation} w_{i} = \left \{ @@ -229,137 +334,25 @@ w_{i} = \left \{ \notag \end{equation} -where $\alpha, \beta,$ and $\eta$ are constant, z is a random number between [0; d], where d is a time duration, to avoid the case where two sensors to be active at the same time. $l(e_i, r_i)$ is the function computing the lifetime of sensor $s_i$ in terms of its current remaining energy $e_i$ and its sensing range $r_i$. -DESK uses two types of messages, mACTIVATE message by which a sensor informs others that it becomes active, and mASK2SLEEP by which a sensor suggests a neighbor to go to sleep due to its uselessness. The concept of uselessness (or redundancy) of a neighbor refers to one that does not contribute to the perimeter coverage of the considered sensor. This means that the segment of the perimeter of the considered sensor overlapping with that neighbor is already covered by active sensors. +where $\alpha, \beta,$ and $\eta$ are constant, z is a random number between [0; d], where d is a time duration, to avoid the case where two sensors are active at the same time. $l(e_i, r_i)$ is the function computing the lifetime of sensor $s_i$ in terms of its current remaining energy $e_i$ and its sensing range $r_i$. +DESK uses two types of messages, mACTIVATE message by which a sensor informs others that it becomes active, and mASK2SLEEP by which a sensor suggests a neighbor to go to sleep due to its uselessness. The concept of uselessness (or redundancy) of a neighbor means that this neighbor does not contribute to the perimeter coverage of the considered sensor. That is to say that the segment of the perimeter of the considered sensor overlapping with that neighbor is already covered by active sensors. -Typically, the algorithm works as follows. At the beginning of each round, there are no active sensors. All sensors are in listening mode, i.e. all wait for the time to make a decision while still doing sensing job. All the sensor nodes collect the information (coordinates, current residual energy, and sensing range) from the one-hop neighbors. Each sensor stores this information into a list L in the increasing order of the angle $\alpha $. Each sensor sets its timer $w_i$ with the assumption that all of its neighbors need it to join the network. When the sensor node $s_j$ joins the network, it broadcasts a mACTIVATE message to inform all of its one hop neighbors about its status change. Its neighbors execute the perimeter coverage model to recalculate their coverage level. If a node finds any neighbor u that is useless in covering its perimeter, i.e., the perimeter that u covers is covered by other active neighbors, it will send mASK2SLEEP message to that sensor u. When the sensor node receives mASK2SLEEP message, it updates its counter $n_i$, contribution $c_i$ to coverage level, and recalculate waiting time $w_i$. It then -check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e., it receives mASK2SLEEP message from all of its neighbors), then it will send message mGOSLEEP to all of its neighbors telling them that it is about to go to sleep, and set a timer $R_i$ for waking up in next round and at last go to sleep. If sensor node receives a mGOSLEEP message, it removes the neighbor sending that message out of its list L. All the sensors have to decide their status in the decision phase. After that, the active sensors perform the sensing task during the sensing phase. +Typically, the algorithm works as follows. At the beginning of each round, there is no active sensor. All sensors are in listening mode, i.e. they wait for the time to make a decision while still doing sensing job. All the sensor nodes collect the information (coordinates, current residual energy, and sensing range) from the one-hop neighbors. Each sensor stores this information into a list L in the increasing order of the angle $\alpha $. Each sensor sets its timer $w_i$ with the assumption that all of its neighbors need it to join the network. When the sensor node $s_j$ joins the network, it broadcasts a mACTIVATE message to inform all of its one hop neighbors about its status change. Its neighbors execute the perimeter coverage model to recalculate their coverage level. If a node finds any neighbor u that is useless in covering its perimeter, i.e., the perimeter that u covers is covered by other active neighbors, it will send mASK2SLEEP message to that sensor u. When the sensor node receives mASK2SLEEP message, it updates its counter $n_i$, contribution $c_i$ to coverage level, and recalculate waiting time $w_i$. It then +check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e., it receives mASK2SLEEP message from all of its neighbors), then it will send message mGOSLEEP to all of its neighbors informing them that it is about to go to sleep, and set a timer $R_i$ for waking up in next round and at last go to sleep. If a sensor node receives a mGOSLEEP message, it removes the neighbor sending that message out of its list L. All the sensors have to decide their status in the decision phase. After that, the active sensors perform the sensing task during the sensing phase. %The period the average -\begin{table}[h] - -\begin{flushleft} -\centering -\caption{Main characteristics of some coverage approaches in literature.} -\label{x11} - \begin{tabular}{@{} cl*{13}c @{}} - & & \\ - & & \multicolumn{10}{c}{Characteristics} \\[2ex] - \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds or Periods} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ - \cmidrule[1pt]{2-14} - - -& \tiny Z. Abrams et al. (2004)~\cite{ref114} & \OK &\OK & \OK & & & &\OK & \OK & & \OK & & &\\ - -& \tiny M. Cardei and D. Du (2005)~\cite{ref115} & & \OK & & \OK & & & \OK & \OK & & \OK & & &\\ - -& \tiny S. Slijepcevic and M. Potkonjak (2001)~\cite{ref116} & & \OK & \OK & & & & \OK & \OK & & \OK & & &\\ - -& \tiny Manjun and A. K. Pujari (2011)~\cite{ref117} & & \OK & & \OK & & & \OK & & \OK & & & &\\ - -& \tiny M. Yang and J. Liu (2014)~\cite{ref118} & & \OK & \OK & & & & \OK & & \OK & & & & \\ - -& \tiny S. Wang et al. (2010)~\cite{ref144} & & \OK & \OK & & & & \OK & & \OK & & \OK & & \\ - -& \tiny C. Lin et al. (2010)~\cite{ref147} & & \OK & \OK & & & & \OK & & \OK & & & & \\ - -& \tiny S. A. R. Zaidi et al. (2009)~\cite{ref148} & & \OK & \OK & & & & \OK & & \OK & & & & \\ - -& \tiny Y. Li et al. (2011)~\cite{ref142} & & \OK & \OK & & & \OK & \OK & \OK & & \OK & & \OK &\\ - -& \tiny H. M. Ammari and S. K. Das (2012)~\cite{ref152} & \OK & \OK & \OK & & \OK & & \OK & & \OK & & \OK & &\\ - -& \tiny L. Liu et al. (2010)~\cite{ref150} & & \OK & & \OK & & \OK & & \OK & & \OK & & &\\ - -& \tiny H. Cheng et al. (2014)~\cite{ref119} & & \OK & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny M. Rebai et al. (2014)~\cite{ref141} & & \OK & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny L. Aslanyan et al. (2013)~\cite{ref151} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ - -& \tiny X. Liu et al. (2014)~\cite{ref143} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ - -& \tiny F. Castano et al. (2013)~\cite{ref120} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ - -& \tiny A. Rossi et al. (2012)~\cite{ref121} & & \OK & & \OK & & \OK & \OK & & \OK & \OK & & \OK &\\ - -& \tiny K. Deschinkel et al. (2012)~\cite{ref122} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ - - - -& \tiny A. Gallais et al. (2008)~\cite{ref123} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & \OK &\\ - -& \tiny D. Tian and N. D. Georganas (2002)~\cite{ref124} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny F. Ye et al. (2003)~\cite{ref125} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny H. Zhang and J. C. Hou (2005)~\cite{ref126} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny W. B. Heinzelman et al. (2002)~\cite{ref109} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny T. Yardibi and E. Karasan (2010)~\cite{ref127} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny S. K. Prasad and A. Dhawan (2007)~\cite{ref128} & \OK & & & \OK & & & \OK & & \OK & & \OK & &\\ - -& \tiny S. Misra et al. (2011)~\cite{ref97} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny P. Berman et al. (2005)~\cite{ref130} & \OK & \OK & \OK & & & & \OK & & \OK & \OK & &\\ - -& \tiny J. Lu and T. Suda (2003)~\cite{ref131} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - - - -& \tiny J. Cho et al. (2007)~\cite{ref145} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny V. T. Quang and T. Miyoshi (2008)~\cite{ref146} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ - -%\rot{\rlap{Some Proposed Coverage Protocols in previous literatures}} - -& \tiny D. Dong et al. (2012)~\cite{ref149} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny B. Wang et al. (2012)~\cite{ref134} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny Z. Liu et al. (2012)~\cite{ref135} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny L. Zhang et al. (2013)~\cite{ref136} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & &\\ - -& \tiny S. He et al. (2012)~\cite{ref137} & \OK & \OK & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny Y. Xu et al. (2001)~\cite{GAF} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny C. Vu et al. (2006)~\cite{DESK} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ - -& \tiny X. Deng et al. (2012)~\cite{ref160} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny X. Deng et al. (2005)~\cite{ref133} & \OK & & \OK & & \OK & & \OK & & \OK & & & &\\ - -&\textbf{\textcolor{red}{ \tiny DiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ - -&\textbf{\textcolor{red}{ \tiny MuDiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} &\textbf{\textcolor{red}{\OK}} & & \\ - -&\textbf{\textcolor{red}{ \tiny PeCO Protocol (2015)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ - - \cmidrule[1pt]{2-14} - \end{tabular} - \end{flushleft} - - - -\end{table} - \section{Conclusion} \label{ch2:sec:05} This chapter describes some coverage problems in the literature, with their assumptions and proposed solutions. -The coverage is considered as an essential requirement for many applications in WSNs because the better the coverage of an area of interest, the better the sensing measurements of the physical phenomenon. Therefore, many extensive researches have been focused on that problem. These researches aim at designing mechanisms that efficiently manage or schedule the sensors after deployment, since sensor nodes have a limited battery life. -Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead. -%Whatever the case, this would result in a lower lifetime coverage in WSNs. -As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. A such hybrid approach can provide a good quality coverage and prolong the network lifetime. +The coverage is considered as an essential requirement for many applications in WSNs because the better the coverage of an area of interest is, the better the sensing measurements of the physical phenomenon also is. Therefore, many extensive researches have been focused on that problem. These researches aim at designing mechanisms that efficiently manage or schedule the sensors after deployment, since sensor nodes have a limited battery life. +Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On the one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead. +As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. Such an hybrid approach can provide a good quality coverage and prolong the network lifetime. diff --git a/INTRODUCTION.tex b/INTRODUCTION.tex index 715cab3..f591b02 100644 --- a/INTRODUCTION.tex +++ b/INTRODUCTION.tex @@ -19,7 +19,7 @@ Specifically, the energy-efficient protocols proposed in this dissertation focus \section*{2. Motivation of the Dissertation} \addcontentsline{toc}{section}{2. Motivation of the Dissertation } One of the fundamental challenges in Wireless Sensor Networks (WSNs) is the coverage preservation and the extension of the network lifetime continuously and effectively when monitoring a certain area (or region) of interest. Since sensor nodes have limited battery life; since it is impossible to replace batteries, especially in remote and hostile -environments, it is desirable that a WSN should be deployed with high density because spatial redundancy can then be exploited to increase the lifetime of the network. In such a high-density network, if all sensor nodes were to be activated at the same time, the lifetime would be reduced. To extend the lifetime of the network, the main idea is to take advantage of the overlapping sensing regions of some sensor nodes to save energy by turning off some of them during the sensing phase. Obviously, the deactivation of nodes is only relevant if the coverage of the monitored area is not affected. +environments, it is desirable that a WSN should be deployed with high density because spatial redundancy can then be exploited to increase the lifetime of the network. In such a high-density network, if all sensor nodes were activated at the same time, the lifetime would be reduced. To extend the lifetime of the network, the main idea is to take advantage of the overlapping sensing regions of some sensor nodes to save energy by turning off some of them during the sensing phase. Obviously, the deactivation of nodes is only relevant if the coverage of the monitored area is not affected. Although many works on energy-efficient coverage have been introduced, there is still need for a protocol which can schedule sensor nodes in an efficient way with: a minimum number of active sensors and less communication overhead so as to maintain the coverage and extend the network lifetime as long as possible. The main question is how to reduce the redundancy while maintaining a good coverage with minimum energy consumption? @@ -40,13 +40,13 @@ election and sensor activity scheduling based optimization, where the challenges The main contributions in this dissertation concentrate on designing distributed optimization protocols to extend the lifetime of WSNs. We summarize the main contributions of our research as follows: \begin{enumerate} [i)] -\item We devise a framework to schedule nodes to be activated alternatively such that the network lifetime is prolonged while ensuring that a certain level of coverage is preserved. A key idea in our framework is to exploit a spatial-temporal subdivision. On the one hand, the area of interest is divided into several smaller subregions.On the other hand, the timeline is divided into periods of equal length. In each subregion, the sensor nodes will cooperatively choose a leader which will schedule nodes' activities, and this grouping of sensors is similar to typical cluster architecture. +\item We devise a framework to schedule nodes to be activated alternatively such that the network lifetime is prolonged while ensuring that a certain level of coverage is preserved. A key idea in our framework is to exploit a spatial-temporal subdivision. On the one hand, the area of interest is divided into several smaller subregions. On the other hand, the timeline is divided into periods of equal length. In each subregion, the sensor nodes will cooperatively choose a leader which will schedule nodes' activities, and this grouping of sensors is similar to typical cluster architecture. \item We design a protocol, called the Distributed Lifetime Coverage Optimization (DILCO) protocol, which maintains the coverage and improves the lifetime in WSNs. DILCO protocol is presented in chapter 4. It is an extension of our approach introduced in \cite{ref159}. In \cite{ref159}, the protocol is deployed over only two subregions. In DILCO protocol, the area of interest is first divided into subregions using a divide-and-conquer algorithm and an activity scheduling for sensor nodes is then planned by the elected leader in each subregion. In fact, the nodes in a subregion can be seen as a cluster where each node sends sensing data to the cluster head or the sink node. Furthermore, the activities in a subregion/cluster can continue even if another cluster stops due to too many node failures. DiLCO protocol considers periods, where a period starts with a discovery phase to exchange information between sensors of the same subregion, in order to choose in a suitable manner a sensor node (the leader) to carry out the coverage strategy. In each subregion, the activation of the sensors for the sensing phase of the current period is obtained by solving an integer program. The resulting activation vector is broadcasted by the leader to every node of its subregion. \item %We extend our work that explained in chapter 4 and present a generalized framework that can be applied to provide the cover sets of all rounds in each period. -The MuDiLCO protocol for Multiround Distributed Lifetime Coverage Optimization protocol, presented in chapter 5, is an extension of the approach introduced in chapter 4. In DiLCO protocol, the activity scheduling based optimization is planned for each subregion periodically only for one sensing round. Whilst, we study the possibility of dividing the sensing phase into multiple rounds. In fact, we make a multiround optimization, while it was a single round optimization in our previous contribution. +The MuDiLCO protocol for Multiround Distributed Lifetime Coverage Optimization protocol, presented in chapter 5, is an extension of the approach introduced in chapter 4. In DiLCO protocol, the activity scheduling based optimization is planned for each subregion periodically only for one sensing round. Whilst, we study the possibility of dividing the sensing phase into multiple rounds. In fact, we make a multiround optimization, while it was a single round optimization in our previous contribution. The activation of the sensors is planned for many rounds in advance compared with the previous approach. %\item We devise a framework to schedule nodes to be activated alternatively such that the network lifetime is prolonged while ensuring that a certain level of coverage is preserved. A key idea in our framework is to exploit the spatial-temporal subdivision. On the one hand, the area of interest is divided into several smaller subregions and, on the other hand, the timeline is divided into periods of equal length. In each subregion, the sensor nodes will cooperatively choose a leader which will schedule nodes' activities, and this grouping of sensors is similar to typical cluster architecture. diff --git a/Thesis.toc b/Thesis.toc index 2add095..077327b 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -43,10 +43,10 @@ \contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{47}{chapter.2} \contentsline {section}{\numberline {2.1}Introduction}{47}{section.2.1} \contentsline {section}{\numberline {2.2}Centralized Algorithms}{50}{section.2.2} -\contentsline {section}{\numberline {2.3}Distributed Algorithms}{53}{section.2.3} -\contentsline {subsection}{\numberline {2.3.1}Geographical Adaptive Fidelity (GAF)}{54}{subsection.2.3.1} -\contentsline {subsection}{\numberline {2.3.2}Distributed Energy-efficient Scheduling for K-coverage (DESK)}{56}{subsection.2.3.2} -\contentsline {section}{\numberline {2.4}Conclusion}{59}{section.2.4} +\contentsline {section}{\numberline {2.3}Distributed Algorithms}{52}{section.2.3} +\contentsline {subsection}{\numberline {2.3.1}Geographical Adaptive Fidelity (GAF)}{53}{subsection.2.3.1} +\contentsline {subsection}{\numberline {2.3.2}Distributed Energy-efficient Scheduling for K-coverage (DESK)}{55}{subsection.2.3.2} +\contentsline {section}{\numberline {2.4}Conclusion}{58}{section.2.4} \contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{61}{chapter.3} \contentsline {section}{\numberline {3.1}Introduction}{61}{section.3.1} \contentsline {section}{\numberline {3.2}Evaluation Tools}{61}{section.3.2} diff --git a/bib.bib b/bib.bib index cf41848..f6323b1 100644 --- a/bib.bib +++ b/bib.bib @@ -2220,3 +2220,13 @@ ISSN={2153-0025},} organization={ACM} } +@article{ref236, + title={Hilbert mobile beacon for localisation and coverage in sensor networks}, + author={Bahi, Jacques M and Makhoul, Abdallah and Mostefaoui, Ahmed}, + journal={International Journal of Systems Science}, + volume={39}, + number={11}, + pages={1081--1094}, + year={2008}, + publisher={Taylor \& Francis} +} -- 2.39.5 From 6dd75694bbe0c45a3c4d4893c0fffc709b932eef Mon Sep 17 00:00:00 2001 From: ali Date: Tue, 12 May 2015 18:43:45 +0200 Subject: [PATCH 14/16] Update By Ali --- CHAPITRE_02.tex | 247 +++++++++++++++++++++++++---------------------- INTRODUCTION.tex | 12 +-- Thesis.toc | 13 ++- bib.bib | 11 +++ 4 files changed, 153 insertions(+), 130 deletions(-) diff --git a/CHAPITRE_02.tex b/CHAPITRE_02.tex index 29423d6..0010116 100644 --- a/CHAPITRE_02.tex +++ b/CHAPITRE_02.tex @@ -44,9 +44,17 @@ This chapter concentrates only on area coverage and target coverage problems bec This dissertation mainly focuses on the area coverage problem, where the ultimate goal is to choose the minimum number of sensor nodes to cover the whole sensing field. %We have focused mainly on the area coverage problem. Therefore, we represent the sensing area of each sensor node in the sensing field as a set of primary points and then achieving full area coverage by covering all the points in the sensing field. The ultimate goal of the area coverage problem is to choose the minimum number of sensor nodes to cover the whole sensing region and prolonging the lifetime of the WSN. -Many centralized and distributed coverage algorithms for activity scheduling have been proposed in the literature, based on different assumptions and objectives. In centralized algorithms, a central controller makes all decisions and distributes the results to sensor nodes. The centralized algorithms have the advantage of requiring very low processing power from the sensor nodes which have usually limited processing capabilities. On the contrary, the exchange of packets in large WSNs may consume a considerable amount of energy in a centralized approach compared to a distributed one. Moreover, centralized approaches usually suffer from the scalability and reliability problems, making them less competitive than the network size increases. +Many centralized and distributed coverage algorithms for activity scheduling have been proposed in the literature, based on different assumptions and objectives. In centralized algorithms, a central controller (base station) makes all decisions and distributes the results to sensor nodes. The centralized algorithms have the advantage of requiring very low processing power from the sensor nodes (except for the base station) which have usually limited processing capabilities. On the contrary, the exchange of packets in large WSNs may consume a considerable amount of energy in a centralized approach compared to a distributed one. The exchange of packets is between the sensor nodes and the base station. The centralized algorithms ensure nearly or close to optimal solution . They provide less redundant active sensor nodes during monitoring the sensing field. Moreover, centralized approaches usually suffer from the scalability and reliability problems, making them less competitive than the network size increases. -In distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and only informations from neighboring nodes are used for the activity decision. Compared to centralized algorithms, distributed algorithms reduce the energy consumption required for radio communication and detection accuracy whilst the energy consumption for computation is increased. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give an optimal (or near-optimal) solution based only on local informations. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. Table~\ref{Table0:ch2} shows a comparison between centralized coverage algorithms and distributed coverage algorithms. +In distributed algorithms, on the other hand, the decision process is localized in each individual sensor node, and only informations from neighboring nodes are used for the activity decision. Overall, distributed algorithms are more suitable for large-scale networks, but it can not give an optimal (or near-optimal) solution based only on local informations. They provide more redundant active sensor nodes during monitoring the sensing field. The exchange of packets is between the sensor nodes and their neighbors. Distributed algorithms are more robust against sensor failure. Moreover, a recent study conducted in \cite{ref226} concludes that there is a threshold in terms of network size to switch from a distributed to a centralized algorithm. + +%%Table~\ref{Table0:ch2} shows a comparison between centralized coverage algorithms and distributed coverage algorithms. + + +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + +\iffalse \begin{table}[h!] \caption{Centralized Coverage Algorithms vs Distributed Coverage Algorithms} @@ -75,8 +83,12 @@ In distributed algorithms, on the other hand, the decision process is localized \label{Table0:ch2} \end{table} +\fi -In this dissertation, the sensing field is divided into smaller subregions using divide-and-conquer method. The division continues until the distance between two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication and processing, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to predefined priority metrics. The resulting local optimal schedule from optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally an optimal solution, so the solution for the whole sensing field is near-optimal. +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% +%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% + +In this dissertation, the sensing field is divided into smaller subregions using a divide-and-conquer method. The division continues until the distance between two sensors inside the subregion is 3 or 2 hops maximum. This division makes our protocols more scalable for large networks, less energy consumer for communication and processing, and more reliable against network failure. Our proposed protocols are distributed on the sensor nodes of the subregions. The protocols in each subregion work in independent and simultaneous way with the protocols in the other subregions. If the network is disconnected in one subregion, it will not affect the other subregions of the sensing field. There is no fixed sensor node in the subregion for executing the optimization algorithm. Each period of the network lifetime, the sensor nodes in the subregion cooperate in order to select a sensor node to execute the optimization algorithm according to predefined priority metrics. The resulting local optimal schedule from optimization algorithm is applied within the subregion. The elected sensor node sends a packet to every sensor within the subregion to inform him to stay active or sleep during this period. Each optimization algorithm in a subregion provides locally an optimal solution, so the solution for the whole sensing field is near-optimal. Several algorithms to maintain the coverage and maximize the network lifetime were proposed in~\cite{ref113,ref101,ref103,ref105}. Table \ref{x11} summarizes the main characteristics of some coverage approaches in previous literatures. @@ -84,116 +96,6 @@ In this table, the "SET K-COVER" characteristic refers to the maximum number of -\begin{table}[h!] - -\begin{flushleft} -\centering -\caption{Main characteristics of some coverage approaches in literature.} -\label{x11} - \begin{tabular}{@{} cl*{13}c @{}} - & & \\ - & & \multicolumn{10}{c}{Characteristics} \\[2ex] - \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds or Periods} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ - \cmidrule[1pt]{2-14} - - -& \tiny Z. Abrams et al. (2004)~\cite{ref114} & \OK &\OK & \OK & & & &\OK & \OK & & \OK & & &\\ - -& \tiny M. Cardei and D. Du (2005)~\cite{ref115} & & \OK & & \OK & & & \OK & \OK & & \OK & & &\\ - -& \tiny S. Slijepcevic and M. Potkonjak (2001)~\cite{ref116} & & \OK & \OK & & & & \OK & \OK & & \OK & & &\\ - -& \tiny Manjun and A. K. Pujari (2011)~\cite{ref117} & & \OK & & \OK & & & \OK & & \OK & & & &\\ - -& \tiny M. Yang and J. Liu (2014)~\cite{ref118} & & \OK & \OK & & & & \OK & & \OK & & & & \\ - -& \tiny S. Wang et al. (2010)~\cite{ref144} & & \OK & \OK & & & & \OK & & \OK & & \OK & & \\ - -& \tiny C. Lin et al. (2010)~\cite{ref147} & & \OK & \OK & & & & \OK & & \OK & & & & \\ - -& \tiny S. A. R. Zaidi et al. (2009)~\cite{ref148} & & \OK & \OK & & & & \OK & & \OK & & & & \\ - -& \tiny Y. Li et al. (2011)~\cite{ref142} & & \OK & \OK & & & \OK & \OK & \OK & & \OK & & \OK &\\ - -& \tiny H. M. Ammari and S. K. Das (2012)~\cite{ref152} & \OK & \OK & \OK & & \OK & & \OK & & \OK & & \OK & &\\ - -& \tiny L. Liu et al. (2010)~\cite{ref150} & & \OK & & \OK & & \OK & & \OK & & \OK & & &\\ - -& \tiny H. Cheng et al. (2014)~\cite{ref119} & & \OK & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny M. Rebai et al. (2014)~\cite{ref141} & & \OK & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny L. Aslanyan et al. (2013)~\cite{ref151} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ - -& \tiny X. Liu et al. (2014)~\cite{ref143} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ - -& \tiny F. Castano et al. (2013)~\cite{ref120} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ - -& \tiny A. Rossi et al. (2012)~\cite{ref121} & & \OK & & \OK & & \OK & \OK & & \OK & \OK & & \OK &\\ - -& \tiny K. Deschinkel et al. (2012)~\cite{ref122} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ - - - -& \tiny A. Gallais et al. (2008)~\cite{ref123} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & \OK &\\ - -& \tiny D. Tian and N. D. Georganas (2002)~\cite{ref124} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny F. Ye et al. (2003)~\cite{ref125} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny H. Zhang and J. C. Hou (2005)~\cite{ref126} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny W. B. Heinzelman et al. (2002)~\cite{ref109} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny T. Yardibi and E. Karasan (2010)~\cite{ref127} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny S. K. Prasad and A. Dhawan (2007)~\cite{ref128} & \OK & & & \OK & & & \OK & & \OK & & \OK & &\\ - -& \tiny S. Misra et al. (2011)~\cite{ref97} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny P. Berman et al. (2005)~\cite{ref130} & \OK & \OK & \OK & & & & \OK & & \OK & \OK & &\\ - -& \tiny J. Lu and T. Suda (2003)~\cite{ref131} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - - - -& \tiny J. Cho et al. (2007)~\cite{ref145} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny V. T. Quang and T. Miyoshi (2008)~\cite{ref146} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ - -%\rot{\rlap{Some Proposed Coverage Protocols in previous literatures}} - -& \tiny D. Dong et al. (2012)~\cite{ref149} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny B. Wang et al. (2012)~\cite{ref134} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny Z. Liu et al. (2012)~\cite{ref135} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ - -& \tiny L. Zhang et al. (2013)~\cite{ref136} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & &\\ - -& \tiny S. He et al. (2012)~\cite{ref137} & \OK & \OK & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny Y. Xu et al. (2001)~\cite{GAF} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny C. Vu et al. (2006)~\cite{DESK} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ - -& \tiny X. Deng et al. (2012)~\cite{ref160} & \OK & & \OK & & & & \OK & & \OK & & & &\\ - -& \tiny X. Deng et al. (2005)~\cite{ref133} & \OK & & \OK & & \OK & & \OK & & \OK & & & &\\ - -&\textbf{\textcolor{red}{ \tiny DiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ - -&\textbf{\textcolor{red}{ \tiny MuDiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} &\textbf{\textcolor{red}{\OK}} & & \\ - -&\textbf{\textcolor{red}{ \tiny PeCO Protocol (2015)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ - - \cmidrule[1pt]{2-14} - \end{tabular} - \end{flushleft} - - - -\end{table} @@ -239,13 +141,15 @@ X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision p %Their solutions can be translated to distributed protocols to solve the coverage problem. Distributed algorithms typically operate in rounds for a predetermined duration. At the beginning of each round, a sensor exchanges information with its neighbors and makes a decision to either remain turned on or to go to sleep for the round. This decision is basically made on simple greedy criteria like the largest uncovered area \cite{ref130} or maximum uncovered targets \cite{ref131}. -Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increases network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfilling the needed sensing coverage. The authors in~\cite{ref146} define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode. In addition, a smaller number of active sensors is chosen so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. +Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increases network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfilling the needed sensing coverage. +Bahi et al. \cite{ref236,ref237} propose a distributed localisation algorithm and a scheduling method to maintain the coverage and improve the network lifetime. They suggest a mobile beacon to divide the area of interest into unit squares using Hilbert space filling curve method. They exploit the localization phase to construct sets of active nodes. They provide a local activity scheduling approach for the sensor nodes in the region to ensure the area coverage and to prolong the network lifetime. The experiment results show an improvement in the network lifetime due to reducing the energy consumed by the localisation and coverage algorithms. +The authors in~\cite{ref146} define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode. In addition, a smaller number of active sensors is chosen so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is able to build a sparse coverage set in distributed way by means of only connectivity information. This work only considers that the communication range of the sensor is two times smaller than the sensing one. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160} design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disc of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing field completely. Simulations results show that this approach can prolong the lifetime of the network compared with other works. The works presented in~\cite{ref134,ref135,ref136} focus on coverage-aware, distributed energy-efficient, and distributed clustering methods respectively, which aim at extending the network lifetime, while the coverage is ensured. -In this dissertation, we focus in more details on two distributed coverage algorithms: GAF and DESK, because we compared our proposed coverage optimization protocols with them during performance evaluation. GAF algorithm is chose for comparison as competitor because it is famous, fast, and easy to implement, as well as many authors referred to it in many publications. In addition, DESK algorithm is also selected as competitor in the comparison because it is a full distributed coverage approach. +In this dissertation, we focus in more details on two distributed coverage algorithms: GAF and DESK, because we compared our proposed coverage optimization protocols with them during performance evaluation. GAF algorithm is chosen for comparison as a competitor because it is famous and easy to implement, as well as many authors referred to it in many publications. DESK algorithm is also selected as competitor in the comparison because it works into rounds fashion (network lifetime divides into rounds) similar to our approaches, as well as DESK is a full distributed coverage approach. \subsection{Geographical Adaptive Fidelity (GAF)} @@ -344,6 +248,116 @@ check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e %The period the average +\begin{table}[H] + +\begin{flushleft} +\centering +\caption{Main characteristics of some coverage approaches in literature.} +\label{x11} + \begin{tabular}{@{} cl*{13}c @{}} + & & \\ + & & \multicolumn{10}{c}{Characteristics} \\[2ex] + \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds or Periods} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ + \cmidrule[1pt]{2-14} + + +%& \tiny Z. Abrams et al. (2004)~\cite{ref114} & \OK &\OK & \OK & & & &\OK & \OK & & \OK & & &\\ + +%& \tiny M. Cardei and D. Du (2005)~\cite{ref115} & & \OK & & \OK & & & \OK & \OK & & \OK & & &\\ + +& \tiny S. Slijepcevic and M. Potkonjak (2001)~\cite{ref116} & & \OK & \OK & & & & \OK & \OK & & \OK & & &\\ + +& \tiny Manjun and A. K. Pujari (2011)~\cite{ref117} & & \OK & & \OK & & & \OK & & \OK & & & &\\ + +& \tiny M. Yang and J. Liu (2014)~\cite{ref118} & & \OK & \OK & & & & \OK & & \OK & & & & \\ + +& \tiny S. Wang et al. (2010)~\cite{ref144} & & \OK & \OK & & & & \OK & & \OK & & \OK & & \\ + +& \tiny C. Lin et al. (2010)~\cite{ref147} & & \OK & \OK & & & & \OK & & \OK & & & & \\ + +& \tiny S. A. R. Zaidi et al. (2009)~\cite{ref148} & & \OK & \OK & & & & \OK & & \OK & & & & \\ + +& \tiny Y. Li et al. (2011)~\cite{ref142} & & \OK & \OK & & & \OK & \OK & \OK & & \OK & & \OK &\\ + +& \tiny H. M. Ammari and S. K. Das (2012)~\cite{ref152} & \OK & \OK & \OK & & \OK & & \OK & & \OK & & \OK & &\\ + +& \tiny L. Liu et al. (2010)~\cite{ref150} & & \OK & & \OK & & \OK & & \OK & & \OK & & &\\ + +& \tiny H. Cheng et al. (2014)~\cite{ref119} & & \OK & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny M. Rebai et al. (2014)~\cite{ref141} & & \OK & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny L. Aslanyan et al. (2013)~\cite{ref151} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ + +& \tiny X. Liu et al. (2014)~\cite{ref143} & & \OK & \OK & & & & \OK & & \OK & \OK & \OK & &\\ + +& \tiny F. Castano et al. (2013)~\cite{ref120} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ + +& \tiny A. Rossi et al. (2012)~\cite{ref121} & & \OK & & \OK & & \OK & \OK & & \OK & \OK & & \OK &\\ + +& \tiny K. Deschinkel et al. (2012)~\cite{ref122} & & \OK & & \OK & & & \OK & & \OK & \OK & & &\\ + + + +& \tiny A. Gallais et al. (2008)~\cite{ref123} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & \OK &\\ + +& \tiny D. Tian and N. D. Georganas (2002)~\cite{ref124} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny F. Ye et al. (2003)~\cite{ref125} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny H. Zhang and J. C. Hou (2005)~\cite{ref126} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny W. B. Heinzelman et al. (2002)~\cite{ref109} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny T. Yardibi and E. Karasan (2010)~\cite{ref127} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny S. K. Prasad and A. Dhawan (2007)~\cite{ref128} & \OK & & & \OK & & & \OK & & \OK & & \OK & &\\ + +& \tiny S. Misra et al. (2011)~\cite{ref97} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny P. Berman et al. (2005)~\cite{ref130} & \OK & \OK & \OK & & & & \OK & & \OK & \OK & &\\ + +& \tiny J. Lu and T. Suda (2003)~\cite{ref131} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + + + +& \tiny J. Cho et al. (2007)~\cite{ref145} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny V. T. Quang and T. Miyoshi (2008)~\cite{ref146} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ + +%\rot{\rlap{Some Proposed Coverage Protocols in previous literatures}} + +& \tiny D. Dong et al. (2012)~\cite{ref149} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny B. Wang et al. (2012)~\cite{ref134} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny Z. Liu et al. (2012)~\cite{ref135} & \OK & & \OK & & & & \OK & & \OK & & \OK & &\\ + +& \tiny L. Zhang et al. (2013)~\cite{ref136} & \OK & & \OK & & & \OK & \OK & & \OK & & \OK & &\\ + +& \tiny S. He et al. (2012)~\cite{ref137} & \OK & \OK & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny Y. Xu et al. (2001)~\cite{GAF} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny C. Vu et al. (2006)~\cite{DESK} & \OK & & \OK & & \OK & & \OK & & \OK & & \OK & &\\ + +& \tiny X. Deng et al. (2012)~\cite{ref160} & \OK & & \OK & & & & \OK & & \OK & & & &\\ + +& \tiny X. Deng et al. (2005)~\cite{ref133} & \OK & & \OK & & \OK & & \OK & & \OK & & & &\\ + +&\textbf{\textcolor{red}{ \tiny DiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ + +&\textbf{\textcolor{red}{ \tiny MuDiLCO Protocol (2014)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} &\textbf{\textcolor{red}{\OK}} & & \\ + +&\textbf{\textcolor{red}{ \tiny PeCO Protocol (2015)}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & & & \textbf{\textcolor{red}{\OK}} & \textbf{\textcolor{red}{\OK}} & & \textbf{\textcolor{red}{\OK}} & &\textbf{\textcolor{red}{\OK}} & & \\ + + \cmidrule[1pt]{2-14} + \end{tabular} + \end{flushleft} + + + +\end{table} @@ -351,8 +365,7 @@ check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e \label{ch2:sec:05} This chapter describes some coverage problems in the literature, with their assumptions and proposed solutions. The coverage is considered as an essential requirement for many applications in WSNs because the better the coverage of an area of interest is, the better the sensing measurements of the physical phenomenon also is. Therefore, many extensive researches have been focused on that problem. These researches aim at designing mechanisms that efficiently manage or schedule the sensors after deployment, since sensor nodes have a limited battery life. -Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On the one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead. -As shown in table \ref{Table0:ch2}, each of the two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. Such an hybrid approach can provide a good quality coverage and prolong the network lifetime. +Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On the one hand, the full centralized coverage algorithms provide optimal or near optimal solution with low computation power for the sensors (except for the base station) but they deplete the battery power due to the communication overhead, so they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and the communication between neighbors may be large especially for dense networks. Distributed coverage algorithms are reliable and scalable. The two coverage approaches has advantages and disadvantages. Therefore, each approach can be used based on the application requirements. We conclude from this chapter that it is desirable to introduce an hybrid approach to take into account the advantages of both centralized and distributed coverage approaches. Such an hybrid approach can provide a good quality coverage and prolong the network lifetime. diff --git a/INTRODUCTION.tex b/INTRODUCTION.tex index f591b02..941d353 100644 --- a/INTRODUCTION.tex +++ b/INTRODUCTION.tex @@ -10,7 +10,7 @@ The enormous development of wireless networks, with the emergence of fourth and WSN is an ad hoc wireless networks, which consists of a large number of wireless cheap devices called sensors. A sensor node can perform communication, sensing, processing, and storage tasks with a limited capability. It can be used by human to monitor physical phenomena remotely and without any outside intervention. Wireless sensor nodes are self-contained units equipped with a radio transceiver, a microcontroller, a small memory, and a power source, usually a battery. These sensor nodes are cooperating together autonomously to perform the assigned tasks. The distributed self-organization and self-configuration capabilities of wireless sensor nodes enable myriad applications for monitoring, sensing and controlling the physical world. -The sensor nodes have several limitations, such as the power source, processing capability, bandwidth, uncertainty of sensed data, and the vulnerability of sensor nodes to the physical world. These limitations have been tackled by many researchers during the last years, and consequently, many solutions are taking these constraints into account have been proposed. Sensor nodes are battery-powered without means of recharging or replacing, usually due to environmental (hostile or unpractical environments) or cost reasons. %Since batteries are the most important limited resource inside sensor nodes, it is desirable that WSNs are deployed with high densities so as to exploit the overlapping sensing regions of some sensor nodes to save energy by turning off some of them during the sensing phase to prolong the network lifetime. +The sensor nodes have several limitations, such as the power source, processing capability, bandwidth, uncertainty of sensed data, and the vulnerability of sensor nodes to the physical world. These limitations have been tackled by many researchers during the last years, and consequently, many solutions taking these constraints into account have been proposed. Sensor nodes are battery-powered without means of recharging or replacing, usually due to environmental (hostile or unpractical environments) or cost reasons. %Since batteries are the most important limited resource inside sensor nodes, it is desirable that WSNs are deployed with high densities so as to exploit the overlapping sensing regions of some sensor nodes to save energy by turning off some of them during the sensing phase to prolong the network lifetime. Since the network lifetime depends on sensor lifetime, the power depletion represents the most significant part when designing of the WSN protocols due to the limited capacity of the sensor batteries. The major goal is to extend the network lifetime, taking into consideration the energy source limitations. Several energy-efficient approaches have been suggested to minimize the energy consumption and extend the network lifetime during monitoring a certain area by a WSN. %For example, one of the ways is to turn off the redundant sensors and put them in sleep mode to maintain the energy, whilst the active sensors perform the sensing coverage task during their life. Specifically, the energy-efficient protocols proposed in this dissertation focus on the area coverage problem in WSNs. The major goal of the area coverage problem is to ensure monitoring the entire sensing field for as long as possible. The area coverage problem is closely related to the performance of WSNs in many applications, such as monitoring a battlefield, target detection, tracking, personal protection, animal habit monitoring, and homeland security. @@ -23,7 +23,7 @@ environments, it is desirable that a WSN should be deployed with high density be Although many works on energy-efficient coverage have been introduced, there is still need for a protocol which can schedule sensor nodes in an efficient way with: a minimum number of active sensors and less communication overhead so as to maintain the coverage and extend the network lifetime as long as possible. The main question is how to reduce the redundancy while maintaining a good coverage with minimum energy consumption? - +\iffalse \section*{3. The Objective of this Dissertation} \addcontentsline{toc}{section}{3. The Objective of this Dissertation} The primary objective of this dissertation is to develop energy-efficient distributed optimization protocols in wireless sensor networks that optimize both coverage and network lifetime. The developed protocols should schedule node activities (wake up and sleep stages) with the objective of maintaining a good coverage ratio while maximizing the network lifetime. @@ -32,9 +32,9 @@ The proposed protocols should be able to combine two efficient techniques: netwo election and sensor activity scheduling based optimization, where the challenges include how to select the most efficient leader and the best representative active nodes which take the mission of monitoring during the current period. In addition, the developed optimization protocols should be able to perform a distributed optimization process, by subdividing into subregions the region of interest where the sensor nodes in each subregion collaborate to select the leader which execute the optimization algorithm. - +\fi -\section*{4. Main Contributions of this Dissertation} +\section*{3. Main Contributions of this Dissertation} \addcontentsline{toc}{section}{4. Main Contributions of this Dissertation} %The coverage problem in WSNs is becoming more and more important for many applications ranging from military applications such as battlefield surveillance to the civilian applications such as health-care surveillance and habitant monitoring. The main contributions in this dissertation concentrate on designing distributed optimization protocols to extend the lifetime of WSNs. We summarize the main contributions of our research as follows: @@ -50,7 +50,7 @@ The MuDiLCO protocol for Multiround Distributed Lifetime Coverage Optimization p %\item We devise a framework to schedule nodes to be activated alternatively such that the network lifetime is prolonged while ensuring that a certain level of coverage is preserved. A key idea in our framework is to exploit the spatial-temporal subdivision. On the one hand, the area of interest is divided into several smaller subregions and, on the other hand, the timeline is divided into periods of equal length. In each subregion, the sensor nodes will cooperatively choose a leader which will schedule nodes' activities, and this grouping of sensors is similar to typical cluster architecture. -\item We have designed a third protocol, called Perimeter-based Coverage Optimization (PeCO). +\item We design a third protocol, called Perimeter-based Coverage Optimization (PeCO). %which schedules nodes’ activities (wake up and sleep stages) with the objective of maintaining a good coverage ratio while maximizing the network lifetime. This protocol is applied in a distributed way in regular subregions obtained after partitioning the area of interest in a preliminary step. It works in periods and is based on the resolution of an integer program to select the subset of sensors operating in active status for each period. We have proposed a new mathematical optimization model. Instead of trying to cover a set of specified points/targets as in my previous protocols and most of the methods proposed in the literature, we formulate an integer program based on perimeter coverage of each sensor. The model involves integer variables to capture the deviations between the actual level of coverage and the required level. The idea is that an optimal scheduling will be obtained by minimizing a weighted sum of these deviations. This contribution is demonstrated in chapter 6. @@ -66,7 +66,7 @@ We have proposed a new mathematical optimization model. Instead of trying to co % \section{ Refereed Journal and Conference Publications} -\section*{5. Dissertation Outline} +\section*{4. Dissertation Outline} \addcontentsline{toc}{section}{5. Dissertation Outline} The dissertation is organized as follows: the next chapter presents a scientific background about wireless sensor networks. Chapter 2 states a review of the related literatures to the coverage problem in WSNs, prior works and current works. Evaluation tools and optimization solvers are investigated in chapter 3. Chapter 4 describes the proposed DiLCO protocol, while chapter 5 and 6 respectively present the MuDiLCO and PeCO protocols. Finally, we conclude our work in chapter 7. diff --git a/Thesis.toc b/Thesis.toc index 077327b..e58a3a6 100644 --- a/Thesis.toc +++ b/Thesis.toc @@ -9,9 +9,8 @@ \contentsline {chapter}{Introduction }{19}{chapter*.8} \contentsline {section}{1. General Introduction }{19}{section*.9} \contentsline {section}{2. Motivation of the Dissertation }{20}{section*.10} -\contentsline {section}{3. The Objective of this Dissertation}{20}{section*.11} -\contentsline {section}{4. Main Contributions of this Dissertation}{20}{section*.12} -\contentsline {section}{5. Dissertation Outline}{22}{section*.13} +\contentsline {section}{4. Main Contributions of this Dissertation}{20}{section*.11} +\contentsline {section}{5. Dissertation Outline}{21}{section*.12} \contentsline {part}{I\hspace {1em}Scientific Background}{23}{part.1} \contentsline {chapter}{\numberline {1}Wireless Sensor Networks}{25}{chapter.1} \contentsline {section}{\numberline {1.1}Introduction}{25}{section.1.1} @@ -42,11 +41,11 @@ \contentsline {section}{\numberline {1.11}Conclusion}{46}{section.1.11} \contentsline {chapter}{\numberline {2}Related Works on Coverage Problems}{47}{chapter.2} \contentsline {section}{\numberline {2.1}Introduction}{47}{section.2.1} -\contentsline {section}{\numberline {2.2}Centralized Algorithms}{50}{section.2.2} -\contentsline {section}{\numberline {2.3}Distributed Algorithms}{52}{section.2.3} +\contentsline {section}{\numberline {2.2}Centralized Algorithms}{49}{section.2.2} +\contentsline {section}{\numberline {2.3}Distributed Algorithms}{51}{section.2.3} \contentsline {subsection}{\numberline {2.3.1}Geographical Adaptive Fidelity (GAF)}{53}{subsection.2.3.1} \contentsline {subsection}{\numberline {2.3.2}Distributed Energy-efficient Scheduling for K-coverage (DESK)}{55}{subsection.2.3.2} -\contentsline {section}{\numberline {2.4}Conclusion}{58}{section.2.4} +\contentsline {section}{\numberline {2.4}Conclusion}{59}{section.2.4} \contentsline {chapter}{\numberline {3}Evaluation Tools and Optimization Solvers}{61}{chapter.3} \contentsline {section}{\numberline {3.1}Introduction}{61}{section.3.1} \contentsline {section}{\numberline {3.2}Evaluation Tools}{61}{section.3.2} @@ -104,4 +103,4 @@ \contentsline {chapter}{\numberline {7}Conclusion and Perspectives}{133}{chapter.7} \contentsline {section}{\numberline {7.1}Conclusion}{133}{section.7.1} \contentsline {section}{\numberline {7.2}Perspectives}{134}{section.7.2} -\contentsline {part}{Bibliographie}{150}{chapter*.14} +\contentsline {part}{Bibliographie}{150}{chapter*.13} diff --git a/bib.bib b/bib.bib index f6323b1..4191958 100644 --- a/bib.bib +++ b/bib.bib @@ -2221,6 +2221,17 @@ ISSN={2153-0025},} } @article{ref236, + title={Localization and coverage for high density sensor networks}, + author={Bahi, Jacques M and Makhoul, Abdallah and Mostefaoui, Ahmed}, + journal={Computer Communications}, + volume={31}, + number={4}, + pages={770--781}, + year={2008}, + publisher={Elsevier} +} + +@article{ref237, title={Hilbert mobile beacon for localisation and coverage in sensor networks}, author={Bahi, Jacques M and Makhoul, Abdallah and Mostefaoui, Ahmed}, journal={International Journal of Systems Science}, -- 2.39.5 From 9edfdcf87db0146ae33e8ba7bd87fea713043167 Mon Sep 17 00:00:00 2001 From: ali Date: Wed, 13 May 2015 00:12:24 +0200 Subject: [PATCH 15/16] Update by Ali --- CHAPITRE_02.tex | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/CHAPITRE_02.tex b/CHAPITRE_02.tex index 0010116..fdd5ded 100644 --- a/CHAPITRE_02.tex +++ b/CHAPITRE_02.tex @@ -142,7 +142,7 @@ X. Deng et al. \cite{ref133} formulate the area coverage problem as a decision p Distributed algorithms typically operate in rounds for a predetermined duration. At the beginning of each round, a sensor exchanges information with its neighbors and makes a decision to either remain turned on or to go to sleep for the round. This decision is basically made on simple greedy criteria like the largest uncovered area \cite{ref130} or maximum uncovered targets \cite{ref131}. Cho et al.~\cite{ref145} propose a distributed node scheduling protocol, which can retain sensing coverage needed by applications and increases network lifetime via putting in sleep mode some redundant nodes. In this work, the Effective Sensing Area (ESA) concept of a sensor node is used, which refers to the sensing area that is not overlapping with another sensor's sensing area. A sensor node can determine whether it will be active or turned off by computing its ESA. The suggested work permits sensor nodes to be in sleep mode opportunistically whilst fulfilling the needed sensing coverage. -Bahi et al. \cite{ref236,ref237} propose a distributed localisation algorithm and a scheduling method to maintain the coverage and improve the network lifetime. They suggest a mobile beacon to divide the area of interest into unit squares using Hilbert space filling curve method. They exploit the localization phase to construct sets of active nodes. They provide a local activity scheduling approach for the sensor nodes in the region to ensure the area coverage and to prolong the network lifetime. The experiment results show an improvement in the network lifetime due to reducing the energy consumed by the localisation and coverage algorithms. +J. M. Bahi et al. \cite{ref236,ref237} propose a distributed approach which consists of two steps: nodes localization and coverage scheduling. They suggest a mobile beacon to divide the area of interest into unit squares using Hilbert space filling curve method. They exploit the localization phase to construct sets of active nodes. They provide a local activity scheduling approach for the sensor nodes to ensure the area coverage and to prolong the network lifetime. The experiment results show an improvement in the network lifetime using their proposed distributed approach. The authors in~\cite{ref146} define a Maximum Sensing Coverage Region problem (MSCR) in WSNs and then propose a distributed algorithm to solve it. The maximum observed area is fully covered by a minimum active sensors. In this work, the major property is to get rid of the redundant sensors in high-density WSNs and putting them in sleep mode. In addition, a smaller number of active sensors is chosen so as to ensure the full area is k-covered, and all events appearing in that area can be precisely and timely detected. This algorithm minimizes the total energy consumption and increases the network lifetime. The Distributed Adaptive Sleep Scheduling Algorithm (DASSA) \cite{ref127} does not require location information of sensors while maintaining connectivity and satisfying a user-defined coverage target. In DASSA, nodes use residual energy levels and feedback from the sink for scheduling the activity of their neighbors. This feedback mechanism reduces the randomness in scheduling that would otherwise occur due to the absence of location information. A graph theoretical framework for connectivity-based area coverage with configurable coverage granularity is proposed~\cite{ref149}. A new coverage criterion and scheduling approach is proposed based on cycle partition. This method is able to build a sparse coverage set in distributed way by means of only connectivity information. This work only considers that the communication range of the sensor is two times smaller than the sensing one. Shibo et al.~\cite{ref137} express the area coverage problem as a minimum weight submodular set cover problem and propose a Distributed Truncated Greedy Algorithm (DTGA) to solve it. They take advantage from both temporal and spatial correlations between data sensed by different sensors, and leverage prediction, to improve the lifetime. The authors in \cite{ref160} design an energy-efficient approach to area coverage problems by transforming the area coverage problem to the target coverage problem taking into account the intersection points among disks of sensors nodes or between disc of sensor nodes and boundaries. They propose two mechanisms for the converted target coverage problems to produce cover sets covering the sensing field completely. Simulations results show that this approach can prolong the lifetime of the network compared with other works. @@ -260,6 +260,7 @@ check if its $n_i$ is decreased to 0 or not. If $n_i$ of a sensor node is 0 (i.e \multicolumn{2}{c}{\footnotesize Coverage Approach} & \mcrot{1}{l}{50}{\footnotesize Distributed} & \mcrot{1}{l}{50}{\footnotesize Centralized} & \mcrot{1}{l}{50}{ \footnotesize Area coverage} & \mcrot{1}{l}{50}{\footnotesize Target coverage} & \mcrot{1}{l}{50}{\footnotesize k-coverage} & \mcrot{1}{l}{50}{\footnotesize Heterogeneous nodes}& \mcrot{1}{l}{50}{\footnotesize Homogeneous nodes} & \mcrot{1}{l}{50}{\footnotesize Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize Non-Disjoint sets} & \mcrot{1}{l}{50}{\footnotesize SET K-COVER } & \mcrot{1}{l}{50}{\footnotesize Work in Rounds or Periods} & \mcrot{1}{l}{50}{\footnotesize Adjustable Radius} \\ \cmidrule[1pt]{2-14} +& \tiny J. M. Bahi et al. (2008)~\cite{ref236,ref237} & \OK & & \OK & & & & \OK & \OK & & & & &\\ %& \tiny Z. Abrams et al. (2004)~\cite{ref114} & \OK &\OK & \OK & & & &\OK & \OK & & \OK & & &\\ @@ -384,11 +385,6 @@ Many coverage algorithms for maintaining the coverage and improving the network -%Many coverage algorithms for maintaining the coverage and improving the network lifetime in WSNs were proposed. On one hand, the centralized coverage algorithms provide optimal or near optimal solution with low computation power but they deplete the battery power due to the communication overhead, as well as they are not scalable for large size networks. On the other hand, the distributed coverage algorithms provide a lower quality solution in comparison with centralized approaches and consume more power for computation but they are reliable, scalable, and provide low communication overhead on the WSNs. - %none of them ensure the coverage for the sensing field with optimal minimum number of active sensor nodes, and for a long time as possible. In full centralized algorithms, the optimal solutions can be given by using optimization approaches, but in the same time, a high energy is consumed for the execution time of the algorithm and the communications among the sensors in the sensing field. Therefore, the full centralized approaches are not a good candidate to be used especially in large WSNs. Whilst, a fully distributed algorithms can not give optimal solutions because these algorithms use only local information of the neighboring sensors, but in the same time, the energy consumption during the communications and executing the algorithm is highly lower. Whatever the case, this would result in a shorter lifetime coverage in WSNs - - -% Several centralized approaches have been demonstrated, where they are concentrated on modeling the coverage problem and provide the maximum cover set so as to extend the network lifetime. The proposed algorithms are executed in a central node and based on global information. The central node transmits the resulted schedule to other nodes in the network. Even if the centralized algorithms have been produced optimal or near optimal solutions, It seems to be difficult and unpractical to apply the full centralized approaches in WSNs. On the other hand, many distributed algorithms have been described. These approaches seem to be more realistic to be used in WSNs from point of view of designer, but they can not assure optimal or near optimal solutions so as to extend the network lifetime as long time as possible. -- 2.39.5 From e37d5744049f24cc066915783983db40c3ea51c5 Mon Sep 17 00:00:00 2001 From: ali Date: Wed, 13 May 2015 09:49:14 +0200 Subject: [PATCH 16/16] Update by Ali --- CHAPITRE_04.tex | 27 +++++++++++++++------------ Figures/ch4/OneSensingRound.jpg | Bin 0 -> 46748 bytes Thesis.toc | 18 +++++++++--------- 3 files changed, 24 insertions(+), 21 deletions(-) create mode 100644 Figures/ch4/OneSensingRound.jpg diff --git a/CHAPITRE_04.tex b/CHAPITRE_04.tex index 07788db..9770cc7 100644 --- a/CHAPITRE_04.tex +++ b/CHAPITRE_04.tex @@ -58,8 +58,7 @@ There are five possible status for each sensor node in the network: \subsection{Primary Point Coverage Model} \label{ch4:sec:02:02} \indent Instead of working with the coverage area, we consider for each sensor a set of points called primary points. We also assume that the sensing disk defined by a sensor is covered if all the primary points of this sensor are covered. By knowing the position (point center: ($p_x,p_y$)) of a wireless sensor node and it's $R_s$, we calculate the primary points directly based on the proposed model. We use these primary points (that can be increased or decreased if necessary) as references to ensure that the monitored region of interest is covered by the selected set of sensors, instead of using all the points in the area. - -\indent We can calculate the positions of the selected primary +We can calculate the positions of the selected primary points in the circle disk of the sensing range of a wireless sensor node (see figure~\ref{fig1}) as follows:\\ @@ -90,22 +89,26 @@ $X_{23}=( p_x + R_s * (\frac{- 1}{2}), p_y + R_s * (\frac{\sqrt{3}}{2})) $\\ $X_{24}=( p_x + R_s * (\frac{- 1}{2}), p_y + R_s * (\frac{-\sqrt{3}}{2})) $\\ $X_{25}=( p_x + R_s * (\frac{1}{2}), p_y + R_s * (\frac{-\sqrt{3}}{2})) $. -\begin{figure}[h!] + + +\begin{figure} %[h!] \centering - \begin{multicols}{3} + \begin{multicols}{2} \centering -\includegraphics[scale=0.20]{Figures/ch4/fig21.pdf}\\~ ~ ~ ~ ~(a) -\includegraphics[scale=0.20]{Figures/ch4/fig22.pdf}\\~ ~ ~ ~ ~(b) -\includegraphics[scale=0.20]{Figures/ch4/principles13.pdf}\\~ ~ ~ ~ ~(c) -\hfill -\includegraphics[scale=0.20]{Figures/ch4/fig24.pdf}\\~ ~ ~(d) -\includegraphics[scale=0.20]{Figures/ch4/fig25.pdf}\\~ ~ ~(e) -\includegraphics[scale=0.20]{Figures/ch4/fig26.pdf}\\~ ~ ~(f) +\includegraphics[scale=0.33]{Figures/ch4/fig21.pdf}\\~ ~ ~ ~ ~ ~ ~ ~(a) +\includegraphics[scale=0.33]{Figures/ch4/principles13.pdf}\\~ ~ ~ ~ ~ ~(c) +\hfill \hfill +\includegraphics[scale=0.33]{Figures/ch4/fig25.pdf}\\~ ~ ~ ~ ~ ~(e) +\includegraphics[scale=0.33]{Figures/ch4/fig22.pdf}\\~ ~ ~ ~ ~ ~ ~ ~ ~(b) +\hfill \hfill +\includegraphics[scale=0.33]{Figures/ch4/fig24.pdf}\\~ ~ ~ ~ ~ ~ ~(d) +\includegraphics[scale=0.33]{Figures/ch4/fig26.pdf}\\~ ~ ~ ~ ~ ~ ~(f) \end{multicols} \caption{Wireless Sensor Node represented by (a)5, (b)9, (c)13, (d)17, (e)21 and (f)25 primary points respectively} \label{fig1} \end{figure} - + + \subsection{Main Idea} diff --git a/Figures/ch4/OneSensingRound.jpg b/Figures/ch4/OneSensingRound.jpg new file mode 100644 index 0000000000000000000000000000000000000000..8fe77d4e4e4980e9a1941ddf7e08a0c0a295a36d GIT binary patch literal 46748 zcmd?RcT`i~wl*9@MVg=#DMD15iWH^S*Z=_mktQ`FO+Y}J1PF3{_1Oxs+jB(Hn&>`TGiRsV#p+B#~%zxgG96ro^_~?)6(C(&G_&!H}6ya z=Yqnr@`}nT?APj=?@i4ut!+QrJNo(u28V`6M#pC7<`)*1mVd9T61TQ@c1e5W{ewSp zF@c!>CKmAiH^Kfzt`mSo z+Q=%VaF@XMz^|8$UtDokg7`1z0fK;N)DI1g zQUeN#Ua3Qy~(mBY% z?~LYfU~;YDubtW_7GDP{5A0SEk23i3L+ngWs?}Zwo`#6)&Jp7> zOHX#+mUVH>p8r%-Gw4w<7$gVCEPRC5^y*&+UkBBhJV@vyAqkflppy-@_Y%iS8Z}_#HmSurk89Ie}6MoobajkLTap*v+X^OkX zxZc84A+4nH@G;$nb2bV&0o)ZHzW;7rKacw)$CZf;&79i6}A z##XUve_elcbabjGRJ+IG9k51ShhomZ``;hZZB$3feX`YeaAkf3&Up>RK4la(XC%0L ztny{{VfiDrpGG?Oq9HC?kOnPIFFh_>?uCM)Q8mcTlIeCq3$#IuvG=gMoYe=Y!pQme zw;Ypn9(eGY>VOb1MBkL;FSrN;bi{#R5(76Ywp}?0ObmBU+aj%AO==QIyxXZD3gX-C zPN8_we+++`YV9Y*65>*Bcb+yasXXEO;4x^qK#7=+Pnkh$%&Aa$$G*yUNptFXl?Xvzk%Sne4-9&!s|cc6oSZ3l}S6OL|_rPMwe$Da178fY%?+r z0qzZylW?UR&npnW8|;;huE>dG*(%UfqZm*R5re`jO@zr#a~m(hKK}L)zBhml6o)+; zh@If9Q+y@?OV!z~RwRk(rpva5S$M&}>IW*3FuS8c~zX`iMy(OC%K zCN1P_cgn0rXc(l~_vIj7EW;jq<`75?p1}HB2Zkdg?!wm0#@xfiIvc8nC=SDG)vMDA zq@G)To?b3U=@qX%>hXhq(j+a(XJfv!oc4h#A>MzhfY%*;)38((i~)N76wz7}zitzjjz99QU2*Hun1w7h z+g(jyfD8!;l#dgpmB*qy;)+-eh?_ZXyHTm-H-un>ix6_f9nVDw{hoDv>zwFS-SNur z*z(XP@Q7XiGN}qz>urx7?%o@JXGn2M(`@Lb`E=DEz-xE4hUcvCffQw`C{d1jmI2~H z$hoU)=aT|6*Mt1zkuv3#Us+zfxJ$MrmC&;5yk=-lZFD|zY2Nh1De%>Uq4yl|i(?lK zKi?O9y3-YmY~vbnYs3lG88^rhO~NZE_#5jh?DqCItvh5N*N1RqRL^jYWHvgmkb}^0 zNMD__<6qL4SH%1!jS2@K4T>s8bAU82_W^E6_F%kZxbk4bodIG)j8ac$jL+dx4}7Zo ztdWgZkxoXsTO!(>r02h!GlKb5ovgu^nrFCbCx#+he8$a4j^wj*^ru( zOWn2_(OhZ!W1|Vm4v#}g7{On-nYTdC6k#`jZpL}2kKRb?=)g#1)7b$%(EV>cp!Z~~PV=FM zXg-w*d*8uTkoDpD9%fI}Nw^r%U}*a$VZd>>bGI_va`2Y;Ng?^Rr-$UM!^htWV~^TlW@Fr#*pjSiZV%kA~aCiFqHS+ z#!~^N0S=YJyCqgP0d?|&U+h=<5eOk$sk3ajp45w%g}i?8N{>%Im`ibg&=xvb9x=4` zX|G}O5On->>0-d<(5mh9r+PpYtmd$IG)10zf@noO3*kjbkFBV`!|K!9C9<0%JK^|8 zZNkSNhc0E$BF7=5hs6A+fecVIouB;v$VTKWtT|(GM_EN`z3|$Z^{qWDh&Cws6%&Um zm3xkmg@waS=X_Y*RraQ?HxH+$T@VjH^OL4%*Vgyz)r(m|=etvKgu89^e}!kkV^_Z(yql zqVma+_yz-yDb{kuFD_TVeGm?r6&z??c6xfn>x-Hexg=%&6n>g=k;+U=q@IEBYoBL; zTvopiX(X1F#D@A(u-!wG9I|yY#H?Ya56DJ{w$Mt=LP$DQ8p1|jy|a>d+$KKv^m9m1 zBbot<#g$@TA}*}O&dHsCXop&j6T(`HV-3=LX zsv)g>u@c*m9XEe^epI6a@D&xQ=pp*PFf%= zghCPBf|Or4{GxQ$3c*B)8SGeLfC%P2BUTo;7lvQR6Aqzo)6SdGPh{Wrexk}OInJCi z9rKF-S&y9QVTDPV5$qFRRA_|C*ekCpKN-_J(G+m`skqLavOxM7&x~H&3<}H^(B&a> z8?HpaZR8z)L)gW3T>BxmFcb5}@az&w(_@T|&l5opPYzB|Y_GodN8H=$$Uc}f4BgqQ z4Qj~uB}S5nv{b|)SQDLFE`|Z(g_P|&dWtkw+Vhj<_k6O^4FdMTjq_@|`D8=VGVKjD zl(5_k<*PeI*a2d_6#v|uAtE%N($;aa0RHhO-J*gm&x zIJUdLMp3MEUa9*x1M~&=YzObEb&GuCMzbFal$_Go$RY-o@S@(KCBxN-U@Lpid%r4b zi{797XlK7d)IV$b%iBeYxvwhqcI>!LVW8Thjp~n~E1uf6wxKH{-SS=aDaMbNYOpnU zZ>-%^+Cc4$$#@URlbl4DZbtD@rRFFF6;Rez40=L8&qVlUo+0Pw4nZ=@#tXX$z`GrB zrA6<08B80LkzahS7&g0frwE4ledy;^W^L%{j5W9jU~4!z$90 zh_*z6Z{qx&uRhH}%N2vLjspU z%6B5Qv~TPo727@iMzl`!qKFtto=}xtEcua^nxCEQZ9(CH`_u7W@Ef$(bxQNJYk^ue z-DB=U@}Jz%rX$bB#ohRdi<}W-GNve*>|yoV5SyNKV{E^fYTqAKC=MG~Qot$Myq@Bh_@%84V(`YY-8{cb*GpxCvZmvLt1ZFeKRvq`jJXu-p-T%JF2b3F z8fh8eD!^3Ud=iY7K|h0Sf=2Np&Su`RRV~4;guwk;O1Bf==9%RjhZTW2fw}%Ol?+ge zT)6fna?5O|SBRE$W9g-gil7j?x8 zVbBk7359r!auvcR2a*fSk_*l`-p!)27axEVxP-N0<#06;<$nM6sUUz2Z)dDAKo240 z3s&pf6ea34pNIzSa^m5!=c9T7P{aQDG?~OKc-O@&YJR1sJETlDI^UzS;Ak8^tP!3H zpp0Q6ZoT(CxqHP&uBsyDeURe)lti@WC0#@Ohe1}q(aBIF+rDmUjjg>5`sci!BFRX- zcd4d)p;xxX-xFz;LG}TxNRZ4#)SI=6)s}}n8FolCBs`LqM2k$m{O{R$HW9P`fiz$a|Q8XTga7At? zp65IK-xKHAg!Aed32jBF(lP=Qk5Eq%Dnd%?Lkq0+IWs02AK55o+jsBVpL(kBe1g=D zky1&swPmv8{cZTR<8;IB`sSKcflJ%=d6SEDXoxTL-ndLT%VbgTFFEj5tM}z(<0UOM zG*=`8RHZ__;zs4koLLYB_wI~O^T28bD*&*_0C{_`d*<8VvM43qd6#M(<7)Vs($CNq z!SXwD<834hGKOe7s}fOXP0+X*G2iAO(&yUet++2I_Hs2Zo{kE zvQ=Mj8xw{6!qVEqhiC;K-RaV391|H%)aEa(sU{$!BqBVGjD$?C7^*yOmFymAWq*0* z%OzE%nV|&PRD%xIG1=F$DM2q?w46M#9)5AoL)g3S&@=`p!=m$Rkwo*r$F$z;-?P%8 zg6AhCEI)}Nx|FlR@>GNhxr5+I5eds?1Pv0MT#bQ;VQHBSk{&>+R;OPGOVRoIm5?6Y zz(;=WIjPM4^+PJXE2Mby;<9TH=T`xc_MF*I+)O5fFa%(PK8PT$w5JILSk!kf2Iv$! zo*eahr*-1lx#Y`M2?(wh?QFuf)3_^oU5C*aLzkz;%Wi z4pC1Mq2KG-n^Z)vkPGGnc8b=7e@lhhTIy|iTc2>%KXvl;c=ekjn#<2&o4kl`NU$tq z=3S8JnLfB;8T{-JH87uCey>I-wG1m%OWTnZH$YR&qJBF6}=d{GEX(bSk zp7?1IO-}-fBtw8agPM?^%ndgS4JYQ+jfcSuxg|0FKXogOv5|s zHROl~g%tCvLG(9C`!ZoO?kr(1T2bq!p|3|daYwK(oM*dl}v?KA+xgBep$t$cYVsEh~C7@K}J#?Z#%*z1{OUdiIG&E zeA2UC2I#&>=xk1U?mfVNa@RS||IpM!yB~P1%U|hYbfn1Z)q z7OLC+er*|R7G`7DS1*pWrH?rbCti&8UVfq>*h(G2>Wx!ePWvdDgUjNa1j8(P(2ruNJNx~}_%M5?yJA1esAtBYrXKqt)`i5p=aM&WU zt@4Fg5?o$+L2UV_8GkvpZMugL%2BEaTba<;eH!XrpN$cuW@nzh%lKl zSj_?f4(Q%?~LaO-lpP7Haf?x zi2_bl`sT;Et-OQ+A1msXqMj6t$(j`V1Wd(drXAP|JIQAmr4x09>a;2iWYdI{ zmOf1Cg^WZiqJ{lh{UR%0S}D3~6iN86_V6%3KPo>XhvTLXm@rAmnMm-KYx*=v6N+5X zLWhNwjGKghaIhYiT#>9y60>`jcJb=z;U?&p#7NxCJ4h+^yE0iTka7&^qvt!vTFI*9 zL$(pJz`QQYy=B#L^fvy)vPN~c{N?U}#r19d4*fCn0kd)sdas7Pv<%Ml%M zy5n+R-razLHr`JweFavj-HPVk1+9K|T^jCal$< zAC94E&87FETd{#fbA$X^T@+OL7neIG-2w|l2MWQ76hdyCg^cgbtF-oT!)~n81&FSZ zo<6fQ6Uz0;hQV_>-o*_zKb&v#dfS`(Zb(N1F3wRdW)Dm-F%dy@;sg{QQdLS|vCS&j z5ciBN;oL;&M+Ac|?~Oxm&o&B99EGi*D z8EvhXG)Q`?WSEn%CvK8e?%K5R*lW}o4nKAhJL+PZ3Xq+6Q^E~W9AOXte#T^MlY}7a zd(KA%K#otitEWc_H}fAjwF`X zIi>n5J#@D5Srq3Ql%1+1i3=n&;jJ3Zhi}FY_`S>)QhhxUd`MIBR~KuFCpR?fxMvdz z>w_OAS~Wp^3HUq;?9c*nQF8FZP`K767+^M~_kyJJF?!w{W z_p4wggyJd?Ld~F@h(3QZ?~PeopGffNLBzY7yxQ2%A<8xRM^h4TL0%u$isjGC0?4wU z^(JEHZ2tGn#HrIn>~rmw;CUZzk6~@WNQ%LjNmYyDm13_q3yz|OOwnvfIlk6I3F?Xq z?=mWUPgYnaI1H*js`)gAZ7$h0n*ROSQT6#JFLmB9w2+dogq0b#Sr40$FyVbPAJQrIm0w7 zOIOGG>qUnIeCuX~Bg<(?kcTkbOi~%*=M}hCTY`sY zI8UZfxT#QpTn@iX#C->|XsXP|yM1fm`0|Tt3*`R&w!!KzJr8|6oc!RvoUkTN;Utr} z3s-!;M_;S1S&UDraZc56nrTXQamlcH47_O4XlzpR>s;N=a3O0oesap z!V4}Josr#Gxv2O;BdX`8a>*~~O5-`vIO!R%D7~dYn)U&SoZlR&{f;}bQ?G>lRSf6* z?B4HT6W*v<*Kac(40Sv9m^*&9QLT}$kNyM5^mWrIT2xb$Xu8N(B2&ZjZckVref3mS z@e7u-j%GQkOAWCsH;!B?Q9#>MY#~7@p(>q0KZ_^4kU?EVwcvM;Q@zjJJ9657HDArf z-z;C^=#DYpS-}STyMCr#w#aB&`BJO;*F4#$&XtHj*f5>dYL?Tp)#{+Jj%XWJb6^<# z)(2f$JyF_r0J)6{+2sNZ{Y!T$k&6T;1U1I%ZbIoj1enkI5Vj`VHSF7qB_39@VX}p1 z8O$!{N~El$@#Yk5vKU>nS2=9xK5=ksPC+uS>zcIJ<|c+5(C<*6Sb8`rME+w}T%a$M z2l1as`_)7UKxP-IXVI_@2I!tjj8+$}8mro&##QTXS3fGI zHDFnyy0nr#o-ftbERO~REe;*5cyaaK{e9H%E-UwrBUFS6r9}h)aJ+o^ zMFWS2N(ci)LcVB1x>nW?NGog{2Ft%O$iaOtvWAvH)=%Wodyu?lbI3?<&8t4|$}^@E zaW&y{(q|ff+%sQJP%Cc0q|zMjk~Xs)dKY@fy@D%hww`Rujk}7DT_#DVj%}j65B%1A z?BR)3@R2qv-n?HNX_#)ze;}K~dm@q~*w=8D-jj2jeEIVv0C+~}T%5Fwp}jlBJzu%u z-yJUkV`oFVYs#7k-bm+>3fe?$3|UTwj;-ZYY~1t}Ou4)CF12cJZ(to{1V{0uepepirK>)Kv!5=NaB+%vPP3l(7kkReN>7z=*^+*b&aTZr@GquY1 zcJC0^o7)m5&{>OO%qW<244_Tia4{E%kS46?8I@nwT|Pq4$?HVESN9h#UG#4|Jp{Gt zpkpqC26bZ|Zau_T)s5K)YXqm`%elM+D106GpdM_PS|GNhL2q39kyBC3A3d;Iw)&q0 z!`va`C>$$LZs6;CSl|K!-n$IYIs?QKLGSVm{cPzraZKA$K2-}DZds9YN-FMsXLL&K z9xljJ+PfYf*f$n*PBpH*!2gcHcqnC!E}}S$aaz9yc!%@bP9kLPDS6<}`dhUjPccA_+0t|1_5GR67eet;KIhsFGm_v7bNn@{kR zD_TTay{@lawyFqLd@}CfC0mzW*GU?UtcnYZS$tBvJl>F1Gd>Iz!=sch54KzYl%Qz4 z=zkJy_%r-Jvr!up_LSB;i_Ptml9nr%K=o${uX-A=YR5A?hf7Ik zlDKqvgG&=ir00D86TKQ{orjVopf8m$g3z_2+bv$CwG~-Omxij=WxqD?Z`zfxl{WSi zircB^beU9C`Q%!lcbLp4?9dERFu%;`j@@vO9Yr>QS~G1g^ZwwXJ*eH|&+<#p zDoxU(Jy*nyZqCKPO$hk5D9#;zf7{mdy*qU}&H`%eY+1xMrS!oqTv3O^3r zMmgQVc*=$HsA*hRyTz~8ej7h+Z11&(Nh0-R3ark(29#iMfAD`2Y;8dMKQd6XRCxRR zdA~X54cg}9SjRXf)IOjp`$vM<6%V_ClT}mn0|qG4NNH_^!hI|9Bbz4R|M-Y1|A|(; zBx0y1G2exT<1o7~g6FCzfjj9uSEg;9{HvW3cMxy2cjm*}Y2~raooO9CqZ_+5Snafa z+qK8Axc*Ve&$6DiC1L4$2z#HBv|lv7KQZsm>hlq&TNt1bHv@7cMH{9u)A}(816Vg% zQ-$PdJ9)iad*<3U>&t>Q>cz@biWt372)e#U)X|nAKbdbndcdxRo;Rd?b~snYaXLKQ z7&f2Lxw;P>T$J$|r2(h<;ckvpoFR8ILUeG)7R!P=_)z$|YhHUzjvAZ4? z<$OKglZ1JSP_?^-oWV?U-P7+NdtBAdamcfmT2BFN99llzybIX)S|Yd74!?qJn+R<{ z?u7Lz0C-peVZecF7V?6GIt)O)!l@I4sT7&@SF0zcJp3wmqPmt0`3U#fKE&hBd~4JS zR$6rlc+9)@AS~VuGBHp&+Zoz2jn}X>vE3h$K|y3y@f%O_q4AWXtpZVvab}<29lgYzKN+oa_HWIF2IrYR z{ke1ZZ_U~zX`>huS?lAKoGuvqX@ak{P781Om{=e@0tr?{Q()B{O zw}&Gy`bM0Scj8?=&lG$3!rlYZTmRLZ90PLy>(1YrO+GvoQyyAa_>N0eZ8{-vIVO-7 z)bgB{QhGl3znb-RK#G6e(fO-6zxV2iOiw>QM`%ZuV^o|#i#jmIMm908XgxjxSL==i zkcKqCA!E4hzb*F85&Gv1kQQJ?9XuGI6^dywL)UU*fC9`@XLG&)^*>De;{fT+P0Q2dqnt;KHHZ?GKN6TUjY@Q8 z?_r87F=YLmu6GIg8?#$*+&LXp_OlbnePrvKa#-(nT6bZ=dT&_lXUb|H?F zVu(I10vpt~}II%#kmdAfb!en3QkcAX zua1y_oehomzc!?p9>LzkR?5yBKA(K0;~GcDJJqpolK7(Elt+8x7vU?lTeNJjoKjAP zgZjrsNLk(GI4Lzp=hx*YzaVpGAO{r~Dh@(!*cgTH;+Dn!Lh4^;KFApRg0vm z#5v8=pXU`GU9VR!^&nd9J(|EcS8TcIAVg+3)bsK3e7yNx52r8+cl6uG$GqI->_bJ` z{)c%+umRKNL^kfEv#~H}hWGuKJj=A3vAAK8;Ibl^M3#ck*<^)*>_@zD*T&w{BRbv2$>36zNW zHF`M3b!W!-R}SP9UhFvDxAdf}oN~ZLVQvDusmHwVjdSlogW<$?Y<%1`fCWVA`IKe-iI!il>PKu>)oC1@yX>koKR6R0;Xf~IRmkyh^htT=5 zW_EZrYVJkk`@kXuCC)DBX7hK)*SKqz{6!szykrVPbZ z^*8zj*S^bg4&JTTduZ*ynoDx8tO>~h3KIUUNx;zg>$a53h<-4e5(6ZHXae)p*-J%f zin&NEmCqofEQh8a)nZRapGh9Ls zBKey0KvmnvRWmZc{Y=>LJmZMFAKNa|I;6${lh7C3mzz1YZb*@ZI{-Cg=cu@V?1l6I9`+XHHU;2wTa)y>H%KUwWI|EGd-TT}(U&@c%Tfnpj=H2dB{B9XjB)*Y? zt#!@x45-WL*;HG33}Ig8CGyHwNEio9OQ!_##ATVMm}XV><`)h^-cnqMnWRZ#`Zoqh z&}bHX3eLP6D*c;!3tsfKyy-nfWvSkr5Vj8?hk$8!HmTJ(N=Ab(Eeh$id6fQ3o3osj z2zjIhrHL&@Hj1iLzX~ojj4y3ejHlZ3*(aQaCw31hYb(G7|K7b#NPX&#$KN`qdk&hg zOfSM-sLo~8((*Z@{N77ZU)6!PIwvV8sM-E@^p@s)_%=&KGrrC_E!lnuMKJ3{-9R?| z0(3GGx!47Gzkz?c8r<><1RsH)_ypG{qu-HvtgF9944V(|3rEXX-;Bvu8@tENCDM2x zv`>{gMSP;bXToK8V$;|_dNRNE3Dm7t+kI}W<kI)C^6TEycFD|~`N1LSeCcms) zm^9E=^2Uz*_G{I(#98KyDVFr_djtW@;11;A{q(*+a!(2T59h&{@2HnZkhTi^aX*!H z#ujjJw!h2V{f$B;(GnjfgMY9xad`aT5e|7CGBlJzQA6}FKz#E=NCce|mf14YaE>r= z<@PUfvW6A5=H6&(>g_XpokxV<#Cfy0FAFxQ&U;i?7{K1tZg1rlOc~^q`P$WtdQPIl z+1>!rD|u2jvm(Pbn6~T)?1kQ7fC>SsHeH1Q@&_W#KV4hB%Kx^>ie1kE3^L(8Fu9&3 zSkh&mWj&IQ987S${K-19GSOoM3_@^GFA$C<&4b5OT0#z{$q`K2nX?QKdlMImLkr$6 z`lvwQLR-f*RpBpBzW1n3TV1;8#3x|XUg&Yb+$wxpcJqT(=#Ei=Md{tjoRx-5DL3** zpb+>sa`y&qYYFgZ9*}>$*h0nPfUU|G0X^YW5D_{%3@8S;OXwQ>hOMm}oVBR^lK$!R zJGQ1L<}(~8zserhvzYIBhAP8+ha4wcnUbRm$#lcSlkS%%H~J*XU>`KZ#0Mvv}~lOv=1MQN1x34iR=|nr2Mr%5U)ISenIR~fmpC_tND(P zOl%Q+l=2QGL}sE~U72Djttaad-EJj92}f(tM+@>M^_DKNQ5B;avPbnOoe`&LPmEdy zuVe*+@gYggj%(7b?y5FjEt3-b#lse4hsoETg?`2#@5YbeUD0HZ$_oro6CLm={m8%K zoDc+1F#uBG{y-{D_#Pz`(Fjx=^?trVG@SLCUz_IGT|W8npjl=~_z)Fc_Y*>FSC~TNAGNFGRHNso2qNLu${5{K)U-F|RFiFw~yjL^5)kX5&MLqG(NeRLCYB z4Oj;NWM(nlA<`dd_U#PN@-W>3Urt}hk z_1*?3KFRd`$N}KVPBLyAqWe$d|ECUcYQO*SYq$ykNz;kQB}5$Z-_HQTzgO>|HiZG| z!B8`xOJFkqRm;sWKtE{$fLjF0QU*UUK!0IT#eeRFy8Iz=H%N*BLP7TxaX+>5Bk52e znSOh48Sn=b0B9Kf7tnb4Fa0R(_vnCEH%u5HZQPbS@=cft7xfPXqCmpVu;7Uw=mQXD z4E-FOk8nC9GFI_klcxhy%}}$-iNK``@J}a;n6{K#IIfvhMvQHBV#t^Hx>A3p1)<8F zUv+VWOGZHX`JC@RHhKY!J0X~9Ig6q;y2r+k{cCdkeZOi?j1SQom>6D|e_dYd5 z^U+Z@)C2jeVw+)}I%FU>I3h`Q^d;`i@V!(wY0j49162IpEAQGzNOa<-xe*OwW3%WWXK;alY5%f zPp&?(iFn?7O=9A<`EyJ2SIiZ!tTg2*&HfI0gR1trcLOr?^;ee*V=K^`o^I7U4A7W^ zxIylPHXiO>FXJpj`?_X*1cpH>Q`oZj_MDfS8Tq;^v{YUbB z-av_KKcs{cq|<09oGX+K)<4@?3QXvHP`q~?WG4IgjBRN@zBxT3Zh>p^M(x?-%h$dZ z9nu#7DO{kTzEevKgd|SY?V90aQa8q{hE9=|qo(?}1C#wyZ;iGlF9x9Tr%^J=@q6`& z{(wmVwl-LUT$}`O(raWBV6m%{kgW|M?ip-rv5+3nvT+Yh4y}!0)=THX0y=YY$89Pn z+ajs%N6M9UfW4A5qeq9{cV88+vVC~=#?eCD5r~`5=(Z&-b=PF<#=LRCyo8Zl^U?2y zQ1&kOv>OW-xy#$yC*>fj0;f54P2^_3wi@-i=kv}7Pi;L($eIkf@-YYY>~1cvuc_~A zHwnc8)v=7a-G$y3kAACh;Aa;9ZeDN9%Z-N4%Ei;!Js_e;mqbxdq4>neLFBABVK{m= zI*?tWlSi5SG(O#1;QV1PtFxUKuFZV8rZuqQ<;AWq9>=dbf7kj=76`PeUy`M*-$X8P z5i(M(Rh;NI3BwF}cKX_saD}dCb~4Mg`s4Z5*B>Q9HIzj9jTG1WC&lppXoq)! zq`mQhm*|E#j+V2KFqvzr_=^1b*6+6qg_#x{SAVmEOniqL3?pZ(TGujT*9?&_axNy@ z%RC+UXu@TeRw1GC?EEOD?mg$W(~F!`T}TA-qy!-pem!=U^mXhI4euN9HvY{%30X2; zzDcbQAr4j08)ug*PMKh3Yr{(FHQq2lB@=4_Bu0= zt+;Rp_QV#ta7|+1g?Ls2qe(d9$~W522&0#6ZZ= zRnMgf+s5Prr8UmuRe#31i05c8<~wtFSj|laI<^&+U0FCM*>k=@6D(V&}0C$IiC(_pXs5v zBBr&~X|ENt<>rouQ#6|9O%AUT@<*u_`Q)$Pl5S6mr9sN{7dPnq6F2-+#rU@PLejrM z6K#B6DvtNO#PL@=+}&ha>#YzfNeu+Q`65kQIT5CjdV0azPJVe+)vftw$GynNu0l#V z*~72J53(tq>>FlZf*m20Cm#2^4Uv>)j#$1!Uh(glPP7vAw>Ju~S`4*0c!MKygK0Ka zf6!7Ru#hbOKwVk&K(W*7tObalA+Ld7iJY=;$G{babGbgQolto*!~vkP^}nOC54c-J z8LkyBofqgQ%I$@o3~I;ilufbDaE!B@HWucB;Wh`1lxjS4R`NfZIPLiV-#L)|HahD# zMVu;0D^$V=gJTbtql$MnS;;;WBdaF=#?yVT+8#^7a?Q31zI|IrxQfO^1H~`c^|g6m z|L6{_1n0CV3~Tj*G^xKYC+bB~gE%o`{W9Dqb%#1WD@U6~?KerMR9mwtZwhA>L#jwI z6LsUds~3%BD!oc@_KDgUPO`t=71)JY=iEF2l`gAP0}nzhoYLvicVM??@{iQm$A1<> zx*GwZw?hfWs8OgsA&4r=={aw)xEHIbA({`SidhNWeASj}-vK9oblg^9UYai`2_kvW zl9A5fSR|{qG~yRsX;@qQ#;hFk>NHWZqHno>*fg!VWO4gtlA@6^+p_1m=Ufib`g8s< zfo?{F@iV!Lgq4o<&G(%qVuX!PM;;8Z<(H?##MVQDJM0Sd_A*vO|7Asi6!9M^96o@5 zYQv0BX2eltVJLhOCo|T1Z@>!e(*dQt8BxJMd4cDb7y5QcvbU-LTKFI-+uQ z$)>>X4;xFF@fCxLII$0j$H|_Srip_SwxsB7gTlpt;-$Rc73co-p`UyI#6JG(6t~gO zAiAarM$plwwl7pp@-`7Q7g0S|q!+KLUksCNZ55F-PQHG;)w)k#KSB8}-THOTV`f*S zR^l*~r)qlE3ZFRxewrgmM7NgCOoQ!h^*@qR4C7^%Glp(N3izHA|9oqC>27iSv#*7H zEpB(SoP?sc^>x=$^Q6i`0*89s+zNQNxHy+jU2v4*YrmM}7x=N~SP_=J(^fJ&J3He^ zC4Lj~e=(>G@;1p5pgx?*w(}eL6zNq6UxN%RsF`3d6OM^iYVP+|gl&q)3yILpJuh5$ zzT+om+P!>jn+sq^Jiw!Et+=xMR(q3~K>%x|f<5N3f6s5xY z?7_jI&lJypR_1@@C5h8yDZOX=MB7l=?L8zRol{n0T1tj=g%Un$I%11brMOn4No&Tf z=%H-V7W=fkJXe4$=PfVUrzzx3u`zfRZ+JX2d)Y!-;z|GVPKB{SUQJ`u9s-z=vUcYFx zy{JGY!+lq8$35SWgT&qAbERFl6@5CX%HnmxF|p4DMO#gDRv>&F(L&8~kRav5c=i&- z9G;{!5oTzOpRMkL$7f&XR5kDT?7PsHCul~`VczcyGu_i}E!^pC7d}*+>&&fpM<+~@ zU06>6Hj2hh(}Ef)JEAW;)z%gNux0%Uc&)C&ga5OoCT z=3AKQffiAeLb3vWOO=f)-^im_tT-q)^x+5e^G9Kq(bv8vS;V$Ic5^-_=>U1sK_%$f zM3S^Kf+a{lK-nP1)$3!uEo$u~2gql2AZ69^F^ z7JrJEGN%G;9vr6v)MgxFbQZ(9Og}+Wp=eJ*$`-hhu%zO|M!K>GnqW1LYJrT-x3Lnz zGkYH2~cNQK)|wyZ_-$O4<1N?-D|oK8`L3v`vzm%`an-xPwd4Oy0h05O@`okH8?_mhjzXAYf8SSBECW;#*@{ZM^4p(kNi6b6oUJ`l)+_Qz z2H#c8EO6(3yvm6Rcvspy+UaFme`D)d&(x=lxzv60l-2cJOHX^!+9&CkMC+LMf}eZi zl`J4`mbs`+F+M@5&ZD~~8FG`A*jmF)N7MmP$@~kkFQ{>p6E_Z z8EH00Oy~$+vamp0XIGoTr2D67GC<$r74DGBDeQ=T$PonCMn7~e3LX^tIbOLdXu|HB zOZcEE%g@Dz&MV=9;shpf=B-?yq!6e$^$8^E8IcEliT2PIb=`ymi-fW%aR1L(*X9%+ zu5x@z@+~)xZ{_#nvWQ471O)C(HYG~>#Nv)?pMgH4pZ%Jn?~PfvHn;P6b!+7kY-=Gt z>*(dO*Wz61&-8}n>0PFK)pj0@^=fFf3sMajlMuW;fb|Q5E9UVo7kvu^X0yzF+Iz0G zR}5dQkO;d}QtOWS1tsX>K;$AjTInPDnUR%k+1ObvNz%N-K=zYH@3QfBj)xn%%nFz9 zS09d=*wMvUlf~v};5nA{af&N>v#ET)22oiIoxa!E)5Gy;rMH^7@Qar8A|@6|xPuv& zo=<=2Opt0uC1ivTrKmwX%1X;Bzu2Z;G;h058KZE?ndi;KYpKI=Pjw;R?biL~?` zyrBX46j7LReA+TYE$|qCde%4UhOb`Fnz(V+?x>dh@!$`t6Y6>EN{b6wJ}#kMpC&P$ z(ho}lTu*CcxmpxDxvdXc{@!XdFuthxX=zjY=6H&a!_DNfVgD5_vY76(A{e_JIhtY& zYhW#g^SmQVE)7xce2qib|D1C67Z^y27Cg!GRrc1EGY~j4@pPQ1?ql^T&jExgeH>B= z!Gm*KW;f%If`Kw-MzM6qFHM};75R_v_uY7ulv{l4h;{#0JYc0@>gSEecGO5e3_}Auip_xY z(xVkX)=?OGl1CuXD0CU{Of&g_`lvjGOXQJ-ze7)Jx99yHgDb0hVW?v2&N)VPw@xZf zkXka?-nm3eolf>_*S-ihCvLX(fIYxGbSY2VSjbZ0@8P{^ml&{OuBQ++>Eef3Q$05I zkcFW)t0N(Ok!3!UF$o1SnL)kvn^@!ZHmPq*Z|wR7qncaUl=1{+tvifVzsr{!XIU1i z)bE5?Z}<-$BvE)gr?km*xDh=2qP<-4mDZ z4sPeKGf_Z*d*W6^l0D?cVxOcCruBW4ySwKvNYfUBPtL`hbpe!;2%Q8$uHW@r6PRNo%<`r!pu>7VL5837GuW787tgX6 zf;?B-&NPqCf6{&SzK3iVcblo=vZ&1Rqsp?r)fu#|k)M9fJoyDd0(AtrkJ&Q`<7etL zaZV2C1_X7(rAefNuUm zfvxYvC=3>z5`ARate_p9c?3C1B_9t8Z+DcgqX$OgID;QxCNmE9!A6pdHq}ZlvoA~g z1JEF}drv%sr%tm$fF>NrY0NL`mKz1KT%#sYUitp7%KbKR?;7bUjLrI5pYZ&%*Ak8) zODv_-x9zhtwjIVs>uVHsT%O-<(Nd0FLxsFAVO9Evz?K3gp-dpHwKStW0*s~?^d87T_ng=$_` z-_K`tJ?sR%{aj1VlrvV;f#)S@({on|b(tORor$ZiHEPoz2PIn{PykPVV@Va?3gr8y z%&dSusYk+@rvQ(nN{x)$+nq&zz;YISGp?B?ldvd{I7Hgonf!voyz+sV6M zJZ5hQmd<$U>?>&`z1#+sRZBvyE4NLI^OvRHr(E%qPBYKrvtFq-peK!CRXjRI{cm`* zUAQ&h*$=p%pNZ^}dZ=YtRV_0VFdLf##Yk%bQa3g3k=Dh~;1qb5B=e4@$akKmCGLgu zhGw@oKOd|G8B*+hvs}G-9)HY9K-PxPy`RGb8%`mu2HUFXysd$@c}=nUv*+(nMYx{D zUB2G<<@*nnU@cvuTONu8uBeL28+}j%Yr)D5eIGlC2l}YGs|=!r3~orLb7@JZ){M+p z$2MCwTh7Lw$PrN!=~(N#?YKEDm|l&`>l=NaU2v%VmE+H1Snkr1ypM2i5i7zZjoL8{skkF{a5SH? zf{Wy-s5_9kq0raZ*O@BA(+9M@JCme24>|p)6HHZLZxN3=O`8Rv#?;X`@1|t@w;Gqt zrU1^G)a=%W;v}2%8c)*2r*^@t5YBkRi5l8BY?^?lS?_tAn z8vZ%ko2mmo6m;$P>VZmA<7vc_rzcfedhNm&%e@7-A3c%SBYNWKMg^1@?gi51IF<^I z!7=kh5DWWapal&hM&7K9eVz zvcl5YDdedXs~x5HK;v9^6Zpuo0?;5*A2QT)g_%`k)2tk=tCsUkNpTx9n?#;&mS;bq zv%Y{1dM#}0Rn$39SNWOiEh(AGbX2K&R)L3rsMpGhn4%PTifdxe<3A|igwh-U0sEelmSyzj(M!e;bW2~ z9iyn-Cb#>Ls{4l`T(T-ZSG-Elyc3r5%>^ZXN{mlaYoE_Uh!Z$IgxC%vX#m_=%ih>unz29hr;myy^EW>cN4&$6`*N0IlZ$zi|lXf*mehXNL zeHb=MFruMfG>Cj@xR7$y5NHoA**upQDW8$BZFcGXRP5xD%5X0QizE4m2?A_`U0@3W z1Sm1K(t6^|a3BV~`h9#y{_8zXdwGm!Z-QC5xZF*ZAlZYT)t*dxV<5)?WX*AYCy9rN(syqp-U&i)vF3?u86#y zQnGk>+33}a**)AG0gMogw8e%hd` zyL;02*aPjR*f^{Z>Ja=k8~|9c5y3e>Z`67nm%s^59N7CU`f_haTz6dOqr_L?-EP2M zJ7LyK&gizg{Mx$tOwEX`(vF=MN$eL_v8X-5_(Fl)!=NOY<(>{$Lbsk5N=v;4lB5kcYj@@{oG~Wb)oBBN3}%lcPWX@!oRI@ zxZ#uVN{4KiqF8~w;DYI?^MUP*6YH~@y_6NFrBe^>E<9N?>&vA398A<1_Pol)|Ad2QT zwh_Fh4M-hB5ZE5}#yA{o#gU{YVeMM?yrZ#4ag}$L8E+6G(&}@~K76NMSY2YeiZNDB zPNd(pKIIhZ_kO(j!-!ECJlN_w?17HW%`1%)YdXZd@aCcX=Y`LUz1%Ca2hRvOm-HdF z9yvyZr@4yQ8-PIJ&F3DP-QLPgc%7%^r{PSB9yt>@RyDH2P3NU$SOUR zhN$Lc=WK`L=c4U8ovV`GW(fag@N>`M5xL84@;VdH7P178?rUaN)9`J2XQ~HuHN0~d z)T?SFntyR++ch7a=aKvU{0ml<+;vgU>9Szl%K#_|I%!o&cO_0H=TAaaTh)yV;UCd% zc1mA9+yW^WNI~R}b!JwABM1+AVjFmwPl}MYX+5#1ZXA43Wr0W@VXV%+u-C;(wV6Hq z0!}i}xoaKH!42-<%C~~+6C-|c)lK^|*nw+$z4Re4%VcysUv@~!iiAF5nAG-jsJ0&( z^^FRj!!TuK<524E1Q8+-?(uD)m&5j2G}3RT1-00=dT-fIR2?$h`=pV)bx2EO4fh1{ z8#0u2mB}D~LGArCCWYK0_7$voX7#9i4Fz2pjOUsG%?XtPn~aV}3*>3f~@S{vUFT^OB<2ZCv~9#u(& zdLqEbv_Ic;+)}*~Z@t{RO6sl^hb@~>FcF!@nAKsq$g_jGi)u4#<7(fm0!4MQ;?~#C zZg2$FrY133ITx+BSKr*vifbFd@pz;rA`Em4c7-?R5&)bVlqO>PMpsB%;KPDuZhBrV z-g-BrYLn+|_UG?kvZK#({p)8{?>FRrx@IDoh@Yzwr2O@1?<;LUDQC(qVAS8enU%#g~~_Z5ZaD zTK$;iZHy3N6ZCkH2holh5bSi8zE2y-%NbT2&SmG)liTInO}B4zLsEf^&uS!!mkF87 ze_p4>(qq<4+S4+M*OgMxJ{gAe(|YKT=n>1ry+NXl`vf*H%iAnekoPUx&;N)<{Qi41 zvhk;ALy|ATgIJxz)g`x_*z^G4}AKykGNx2RpL&gI& z6;%aho*#|;t;+hWL&JbP*gh*x2WN;`ze`|XJ^>c;ASaD+sUa}2#uVWDsjZq7oz(31 zfgFc;dXE5qg_VWPE=8~;gTSprA%@S@u7pu>$AZQe$jd}&ZH?Cjo^86v1m!;q?lF6v zBkYC~U}~`R+0H;kzluOH3uF|B!Yi(QhM%!=uQ$ ziEfJMLkY}7Xgz3Nw)<|@gr+MCim8*`%~7iwjP-m`*D{=6AU;I|1JAogHw@n(3`kqKDC!y$MDo#|WMv(zy+gSkS!&V3+ld{d+@bk4WQgB9M|5 z;7(_EW99vAHlDv@lzcOAF?fHv{*L_(!WbOv3IW*onwoW;i5@rkpCoP$0LVG~X5{0!3=f{lR=sw<~GKEz@^n8y^7djah zPG4IC@c>9DAo%$bX3Nl^3qSK*HmKhB@W7f)U!PFNt%Hhi<(pr`xeQQOX|rr0Va5gx zABA8m=5cyYJWH>}GO_j(t55qD@!U!GN%F}C`?VVUKDRHYNzl$qyNWAwz4 zI(=R^HH3M~P^sOSq8n|>8$lYz!>kYCO5c5iED|n@7WXg}tz+1isIXCf39;V^CNUTg z)1d3-VD1fHzZugD=8q3dBpRms&$pe|3`28xvRGk3Df%@()WD<@YnsZZ9^@rLd;M8P zNl^b(UjGF34QH2o!_l-$hGqk5VOPWKtQt}m78YEAAJnUBrc&bN<>JZrjifV;)Vgdd z!5=>gmJl^m5q5Sq(upA+E2{#C#!cEg*C6u0F}@zW+h)GC||o z5_y<#yP+4z0=^-D)xuP!v#KixmSpg$s*&;f4{Yv)KHY-N*iTOj+9veeZwrd#Vx49=Y}!dMF{NECy_G+7FjvcHhCdH?&B-tJlEO6tq(UZ4oqw?)7}?Vm6NkJ zV9U4!C+o+dZwpj*nDw*?p~5|Nf5u%an!$_x2)3 z+>qsf?$u}d0?=e2W5l_F8`75VDo&gz6>>1j51a}6Zby~ncL_QXWXz9|7D}R~;S2sK zr}0SvaD%yQwk9xb{Vy&6{sKtiUk8bc5<)2qKG?|atT$U3gl8^+0_qgi^80G@Vl31M5ZVLZDI>9*hbGyC^K;_U0Z+8It!ihdbHBlARt zk(Ti?L6r(-KEky_cTTWne{nq_!q^x?wvP2={_YW!0PuzvaU4S&{zHj3GwP=fy9^k4 z3o*||0xy7XkU9U0>l*AC5K(9YI`;|)EuuWw{=c}ga7}2Y1hBSWHyN8i)0!TIJe|VX z*Sz2leU7G(2ojJvI1S4==zw0k1pb>l#{9o?y$QSdf)4ug{p4^Y3hDivNCG!tZv%)x ziI_M#Uh}P`#i8|_j&V(S;9;Zd|25Z(#i+%qklSvKzvq))a7&rF;#1*oK|1ad;mqaC5_e8RFC2mC6fKU5ny~ zTe9qced+H^00(@g;By63aC z=#bWB$xETP2kw_v^%bw^ZYR5opDucOe*Aiect=I(&_GbQPK!x$!VJocFy&^hk|Z*o zo8b7!a2CqeEMx$Af$SHwS-=PZ(#Vdm48Sr^D=%N7jHx5sV9ZKdTif@sKxjZ9K(ZkY zU-&k3dC#WW#_X~o%ZD?R>eeb@H_^VdNi!H4ytvJ;I(fwLnu4l)_UE7{BO)i&56N6h zk9#+e0xJjDD!obJ45vO67}ehZsOz(En=&HUjMLxuB_d6S(FRS<;kw?nmK~Sn8|zvR zTE5X4ryby{mL(+bE@teZ=*TV@L6Sn2d;s#o`yK-s+D8#>JC$8x`EZBDEg6jJ?vpwp zmM-$Z5Ws82KF!RcD2pQlhAy%q%7#Qs_+s9lF0qvlZjtV6)zEUW|9rJt|NAoZIN);5 zL*Nj}poH}{No8in#O}4*12QUOYpc$ufW|#OAOh3$oZMy*m zT6LZ9m&XJD_+xs&&=q&0=$ql3!hm$@oI4H2IS3?Y|L|@r%5i{@ zTx44R;gHeWlGQy=$Q8ihD#8@u7DX@x+p^5xPCXtTT%iA8r~g9q!gqC__lB=5)CcyJ zi%KCU#TFNIprdvSmo)18ESfEg+Lm^Ys;K^elGsjZj-36NjbB@@;kGACVp(l?lWz z*Psz-nl>Qq|KY=Zk=}q#b$ zC94RFDs7j?B~gq<0ATA5JUW>d&_p-b2((2KM8iH!C39qnCo+ozC9^(Ovw8_h9uo#W zyb>)PTh}eb>=r)Ck9IJXPkkNW6dXt&Oi!yT9?^TTS+xt*5p7Su65o2e*wHMdJTTTi zY}_jW?n*RRMGU_E_d;@^`TvZ`Lty{^3(CQN3DZ%fy@vm>0ipmHWM~W|70mn%A`79k zpeu>Hg0LtsQ{#^~2xtrZ^T@7KHfBkjP!3QSp2q&C)S_;po1X&>t+<`D1U?AO8x1H6 z&=&XSA+?=$eekdr+HEW9f4=F0e_(0=dk7h3sc?&_-t1&g3z`=nK}q~JfMBbJ1)LfC zD*imB8y|N|dv`Wv6BFRZ(elH_onP6U+et=RMq~AMp;f?X zGy5X`>$sx-<=>VLaED%9jrZNfK~Z-qOUv;NL}BX`UO``6q3Thtom76r$o}O*MKi}U ztSg)`FmD4l@>GAr`5|uH@*NE~WJ5@CQz?I09K;AdG4xQi``}=i;jEB3W z5O@y;ESOtI$24|pG;z2SZ@#@(;f`Hyy1yr`>jK|!BW%+*RxkLZb>3l|u!2G3{(?YDzxvv zGxcuIz{s8}Nd7-vFsg8j)(}M{3Fb(u9brv{n#fU9UCjx~Rciv?P6w_km&KKoP zsngec+q|7#@=$g9RgIzUd0lZ!$jBmxyAasuQikw|;x38^&P#FXa7xXPlJsG|Yx||{ zzZ(Ad^b&lr2Na}wmDWud+31pvLyy=FTKvu`jBa|5h)vOThq!U;d~P0#9~$iq2w6cK zuH}^ku#ju>O;$OlkzY`#I=3)BY}x0PmUqGn1HF3Rz;nf@ICT%hyfy9+z2Es_Gg zb%q`YH8K4Fy+V6C_TC>ApI5NgG2Wz8&d&Q{pPn}q`ni50;KO&3Xk&9kLzqRP*K)1< znJMD8sXKD7KekzTXtqjM`}>wHhncTe4>7@1!)AccD8!7YC&}zL87e-072}&nmhOzz zE)E+F!7EJZltfIV7l;(pxp-;cnS5fO@ zCJD=Qx)j5ft{0UHC-2*Y3I$-+8H>bAAhFvQILy?&E;`y^l4P_b1QP zcyvof$PFy^Tb`@vA1QtseYPm>xBo52cm3ldz`D5vEp>pL!(5Fs#yha!_i{r|;(-x6 z|MzE3sy-8S<$9@kFyY{B`AO_^810UPfc>B|Eov!`W^GGFNb3Z ztJ`-rOO}q?#eWlH9Idg#JEs&rGFjE6-vL^8l;QFkWrO&+q*1-TKQ^t;nMtgxYz5}M z3CL08)j_1bWR@Ex06c27qwpfY*R!SMSdKwwm592Sh$B>-*bS~EKZZQOl5APc>&#GU zW-SH6H#X;c#IK3uk5zTxw@J#B`Rz*Np+MEM7oG}Er7l980S2_hCh|dD?vG5EX}{M~ zcnp}4n;Ov%ocMBpw^}pSDKuF%=qo$<#o&5nIH}WixM$Sq zqm@sD-klLv_hwacG%;+2XdXkuum#CL%^db(M?H}SVR)vJn^Zhr##!6Z_qUDYRw(JV z#T+gzjF{b%C)=NS9;m~BhI870DkG#_!lX4*o0&vKm|w|UcIiWE6K|X>b%?39D7`CK zHTlf)-YKjABV&g-c$@!$J~@`Spsc zf#+bbi^-{?MOA7A52-T?0R1FPx+YjBQ<)aUFsI?3aD+$L+I<@l{TnQs6W5n-TO)NI zcBe=uKpjJg*+DuK<%l}*ck>Jr0&BSB1DED+(Q6+q3aq^R5cooGD{|VUvsW)tbM`w| zncREfeHi`l3^+nBT8;;(S#J3Oocboj3dpoq->@cRnP@v)U#y5#UHJ8oi0K_6@D>xO$)OB>jym}8P2UAN%pzH!)kyw4z^WJLIMxfdIp z?oj^*mhX2V!)veo!bk# zWp1q4^IyXT!X16-WXi$z)d$9&`$lvGN zw&qEzy0CHS@N7TK&Edk7V?C!Xv}$Kf8z1@F{YfXGDDCQF^BJLK z$UX=hJhIMsQEtN#K$`Y%ieZ$IsZJeJhngG2IB6p9)bFbd9Fy{C{idRO2hB$Q6)fJWGr7O>q}X zH*7af&9Qrt^<#r+OI_2~Y8}cun*03<&j{Aw zokP0SrsFTXm9|{&z$IUVjfPV}%NhWQo22Y+W*zT`AZzu^h}lh%~O!R4~cqIA&Yff%|s;|X(~ddbhcH{k55XE8$N9DLtLzlg)? z&vXkc;&q=I1mmu__+;mA)_hFQ$yZFz9ZR%TnVELQK;a#39D< zYo*J;^Zjvr7m6uLHR*{8k%%m&_6UyrRI4EAy?XHVmHl3f;`GZsKxL#yEpgtyrxg%lr6#mHlf|7lNaI8Z`f=5?EzCd9YuMah3Ky zXELn^Y?3?D_`%~gxAN+#O3TzoJi;?bFWo(DzK#Wy)u9h-2i4@;T1p%G8eb>7PFt|Q|NJ1e&$>Pc%D6>yi}GSGI%toN zgU{EDlA21}MA8#0FAFVNF&+&SR*egpalRJFV}#>dCqHT@&5S-;8$}2uXYdv234orx z&oWkr%{Jg1jE_r-_Z2BlK9#p@apGMY9QZx zX5#6tEW@ZgmZqlX*u&}O0E8l28!d8s^6dGt1+V z>`wIUIs8@fT-HmKn#^a%8if>*vMeQ37fcxFG9bh~(ij(J)>04Yiiv$Ar9_lCD;^#( z;N`UQStK7Sj%Ztf=i-6dM)@dGhl%+RB*jQ!3X}xdwgzbB^aa$av3`GO7j;_wl}^Sx zuLr3&5}RTze{ngR2-vjmfp{h~{486{mk&hPIoho}=k&$#c-;BQQXi`=NnKau4^wtp zik1g|8P^3K8p}P;(WE(+;A4RO!&gVF#QtqI|8RmTzI|SVp$OQG>~W9SapwlDLC=+iUya#I!QsdYn)p$_n&X zAvdx1mwrkeAK!1USX}%zUTw1M;}fHyjt7)!q2Pg9qcH<+m_br?2GQ{D^tX>gu^Cys zGJ`n9O4ISqo{!aQUX>K+=oaJs5W2iUlkN1GHvUR_IIX!s0H&xu``Ss2tNBN5SASlZ z@+5eK_X@KV_kg28FG=(+SUiDB5lUF7$*bwo`W>L#B8il%I`4VfE;8@fnBPHLy@6IA z$+3d8G6I3rB-|9~M0qA=ZPvb9il*g7e=nqZCdLYSyX?zO+8VVEg+FQrCEscutL?6J z=&05};Ct=z63*-HrG@I#P;VK>VH%jUfB-9PrL_DbRmsn>E5(3JKkI3RZSVux9c7Ln zy2*`#+u=q}0UE@!eko6HaaCvAWq2SPD2-B7X~f-qgf$@F9s#DkWx_d6VcUO%S{Qd| z4XZiXAWtae7Z*1+k}!7$D6+%YXaDuup~hjvaq%+XkZJ<+23zeH*A->^j&}}|F@fg{P*c4`>+99Q3`R!^2!_m$Px4p695!IU;;(b z|2{qAH49g#{nyzm^2fp`T~fEByZ-fd|79uuVfwHQZqP=uA9tsqlQ8Q}`Nkfs!|mJ2@}<@fH+0R^@fMT*`xndlGm>o%i5 ziwrCqou9OGx*T8oy{MCGwwz8E=$rRb^xAf*X~2`i;zeBsIG+~#y;MR+YA-3^Br<9c zrL$lz$V48b4ByVc_f?RAl2j9<_{P;Z;g%YfS!-9i7nhzmg%U10eT`C5tHGT@30>P+GHg>rh64JNXFmT0s87 zag8H~8RhNyU+G<9S%>3`_8!=G^zzAY$7#od0NH^m&@X(NSqub6xEAsbBDo?PXA9Jq zFQPYc-RzjE8|YW;@iEi@x-x+L0!>j!V`MX*lb$^Lo!C8Tw|*g(|El+TAYxypYCN>q z5lRC35I!3<{D64iUO2jK5h&UN!1Z$qzt7>u8gjxHZ$chIWOjz6a>?U8w=s;V-AanX zE7UMVFeyx!dLS3VsMO(z;TQl@SWY{$04TEY-Uc|yHmCi~HpRj09Sr9;!dJb@7o5PY z@<9j%fXN`pJG$s-EYqh_nFBFoYrvLk0X#T{ZVuG!j^_c~EsJE%F;oY-d^H+!U-=0{ zxRT5j%sCAlG%pK^KpK5OdyL^mFGemRri-!xfZAfVTnA?`l#C%vydjrPx3dNPD3wt_ zX*_aE04YwlDxhMI1cf*AFA6XDFNf+^)$_Lz^?Wm|%mp=$VKfx>0(^p=nEn$IP|B*C?M=xg&``m6d#2L!x zF)l)jYCXHE=v@aY@_>f;Uo!4~b~{&tUqHH4HTScdbV6<952_|7b&znE{8(@A?ql{| zpvMBn;l2m3P-q1xKQp0%#Dms3!VyC;yNUjaTXlr{!9Ca~Zy&Whxv3seP#oL&;a-x> zEu;0%nw9$o5)$33jr5Vqj&Pkq4S#LT8)2ST^LX%e7X3F7ns}jXqt30661zg3V$fss zKa;!u)4=ZJ2C8?RsRLz}BPK!nJ0`?dBiESS4X2gz0(UWRxG0IC@)dz>M1eigu6IUa zs#Kd1`M$tgs%EZdbGnBJ6_{wwNw#)xDyvzMYAR1z!BMB9T596s^JN&ks**%Z*AeBN zFQZ4-{&j1&m@}p&LBG%tv^=ShAwSSx{Vuo zX94B(fA(^=Kx`p54z~u?!3jXqHeBQY%t9aiFYz{U4we#?^hD-DUaM>OFp%oKir;f2 zv4>BqMSbH>SebLkMjp8TgmaIEX9K}06^22(6|Q2)e8BE?HC*y^nh6aK@HR5hSxlyV zP<%Ty&hovgt%8nIP|}kC_XGWFyK91Uif^@~6i?yjHXVnH74=*F9g9^M2hm=0-c*C4 zd&1}Z6cxs6^he+|GsdF}MZvRwqy2M$<}U8y(Q|Z;9GcQJKw$2M=08DCJO@|0wLtbV zz1y@wK;0|CcPyyC@j_5qBTA%!9^P08|Djg>IW!a=`1p9L^cuUV^`5y3Y9ns?A=(wn zlxkl-0uVqPY)0vPVm8unjb%Zd@M#)4iXA{5@!T}^l2@^!Qb{EFVB71jk}pHY_wWiG zvP^Tg{(7&cyJKZ;!xj?S#VQ;1vU%isAt$x-0z7Se)Otj3$HB;!HD>4c^&d*uKm!QC z0m3ohDaYuE1Wn#ydQWyAH`TEFp{BL53xg&lrH;>t9)4(GoREJw*Tv}I2WdaM-gtAz zocx>T5pcNj+I${MN7mKN%=M!CdUuCBX)Cm_gmZIjX*pD4>u+Ab;-WIZiNqFaZ>chK z|HIXy7ni0r zNh;%+PYOi5v!s0DI=@@WSW3yKpWjZA3)%)Ch=1c@HVgrF2{c=4P*1#rwUE^IY+EPMgdR9#f)PeW*fveaVV99{fnNiw5XWn zf&Ve-^lg{hD%H)Pw-56%I={{J;WiZznk;C!l10+*HN_VbSZ3SBeBZ1qGL(xb`ARoG zuF%4;iE|vNdE#GX7W@}o#mVD7LMAYTa&#k_l)@-!V8ZF-M(k*fg5JKtN|7Z(g%`rt zkLOgpDPzrm&5v>zh5EqzTl>0bLAGyR^jSS~k$q4|I)_J8mi1fsP!^mnXh!=xnTwFa*(c4P?AdBo`fpqkoGi5d}|a4jkJ_AJeOm0JN<6%rg2w!Sgq`$ z8{)0S_&uw%f`hU`wY6*P65PnekuQI}Net7AFdE3gfT{T$aYbOmi6NrtWDa)#y=S5) z5-BwaJ;c7rM9@tP>G?6j%_sB=g@&38SsI+(1Hhqy_>F_uRj7lsD&9e?tF?n^JL{@s!*s_C-);0N%SD$ zn@a6xy2!WqNYNp)@bS*hD?_`x|k9F#+h`KE!T5M~x!&Av&V@&-_a}-&p4m0HJye&2VEXv+kfqOlgIagytTm@kL7-DlbeDvs>)vH1o4X zYV8+SsJFCNT`4%!TQ@vfwX-yi+#QAoopoTG=N1hGN7 zeKHhw)Syab`{w+GWKHB-srv`56!`RR!Npx=|N5ynK* z5f3ZLc%5TW0fPqFxmk6Y#f|=^XI2Zf)B^o04QC1e?3qGAO{+}l)iyHecn3#jgl)>S zqpj-Gm1A%h342KFt8_Pp^SD&1PV)2C{AO_#(MKIs=Re=R^Y~pneD6d%n;*x(Fh7iI z)zb5$*#~LOM$AC!-gnHD7MP&!?HVtvhy7;JTrPtk#UB@poto@UHca`*CR z^+_w=HfO&NlOpYguFHZcpBEfgNDuQXmj~_4>Xh6?&S_mTcXy~FeJH4O2@6iG+z69c z!eIZQ2kmQ61U%?j)FrlZ84dQL0u+rt%nqfX#e-_d&FXp=ft79>yjUx7)BQKY{okGn z`F}Uze?GClaXY$TUQZtQssOi7=;hzfR%E(QbxMUh^#`xi7RM z7VWm>_|KgJKe!^58z%}rnYa~@)8c#~I&rET;aOw5)o4m(fI$Zck%WLoSi+ zMoaMS#oC45SyaehO+Oxe7sNc5w7M)yn}H5HG4RYqU|RyWeRq^8MjgUXbz&}EQSxJb zP%1e#m>bs4)5l+iNr0xovELTtZ#*)2+=oay<^p0)a&+g(> z#7GI576P{53`@!Z;cEt`6)jxSAi%z}Tht(j7OYV(_onZ)U?~;C(-r6!(rC^Q6^(8wK_# zt!W#6mu$t+i!N#REM(!gt4g<(8DDlW9A(shP22a2t6nLtXf*4;rewbbeT!>cf=a>^ zbPIJQRbD&U%I)ae!nbqs$MMjwbFFPf?nMTYxfvQWtHL83%!hyE36h9&Cwiwdk-7lqUX)FLiUUOP3Zs#>u-BYknztS+TT`9J6GpOzQ7kHcVNZ;(LFUF zU4b`niP>(ri_;CDSyLdbm?QI?-)PC0dnKlH0jg_7szYSjJ=D9A)?n?n%Nal4L=Jy> zbam9};NBN_;H$S#{djUkF1S20shBCp(7|w{n=C;s@(=-TT@?}^UjU@)*%-rTq79l- z>+|*YXQwz_14Anwqbm-6aKWPDz|i6oMpcbxVb0e=j{P8B%B@uN+aS$?*1QP&YspuFi-;&<>phSvWTzoN|_v#N_OWM{o<_kIU zyoT_Uu6oGZ$i0dQolo`Gv-l^SRJ^uLjut0;&p#a-X!M{Gp+ASWnu!iH?#>(3P|161 zkaB4{8?mrq2anPnVm!;bB4~@nskRE?(1z2%B-J9A_niUj7K{tW-GF7SitC14~ z`fLj~weR+(Ygi2;Z;CMkuljv7DnWCExG%2dLk0_&#ENvUw88g7Ddl5oWrKEMeAWyA^Z7K71LS0W zn6<5pM+pdkXo0#a=s>n1U-oTKDZftDXdioJFs+v(hw9wLsVT=pDtj8$cN+*3X}`EC zty>Qj-H*}JW8TYWF5ZeqoN%%KN^2A!HqA+XabO5GZus&#?g`oxGCWJoUxLzAdMLQo zBX-R@ct%w?qG--`Ds&dLqdMyRrGT^&xOZW3Cd z5h0kjhtuv6{pDXAsvIo#OB8(Uf3L%c>u228uL{N8F`YE#hjxj_X0s;xE?sIvDptK@ zw>rm}*S<-#Ws0wteMwK)eNs)rN>P7R_ffzM{W(yrUuRe{6Imwgt4uImhf&0wXZav~ zpa<9&@Sh@0LCp|;M{^#(hJk)$ytnc&-fYCe)iwxXz;LcHSfU5w7%IMh+j2;q%b78a;{Hb-JxOznnduq3wJ}A{Feh*`1~`fQ zB@NmHdVwu3M_Q``V+rR|Supaxt{N zQllz!S+}QInJE{BX@rTh&xBLZOB@Co#5tOAf6|0jycMI@`@sWGO3M8NG#P#1{r&2f zuGe?|pZ2~huBr9g77L;%qEw}-REdyHlTK_j5l~bDQnFQ$5+O>Fl8~r02_O=hqQIt! zNXdp!1VSPZKtPllIw1i>N+3ZD(Ug7GK9A>j9`3!b_y3T$^~+kxx4!w!Ip!E+GBa+T z_fkrfA^YRGI!RcJKSUg*&RUKaO1Fn;0V@ID*~v9gSkC}_t!@Qv_0pb}K`k37{VcfMuP81_we|SN{=_i<)j1YLbY5kMN!7lF#N&@XTSi zvVT~x*=)$Y%htPs=^Az<5FD09k&5-4txv>NzOXj#3kf$FV9qeL%!9%=}*k zr$#Lu=9RTvM)|SxRv<=Vn`Skg=UFS)tEra^Kig>uI??puWw)IRE8MH%DNCvf4<7a+j(TV5O<(;9fgWX-3p;d3~J1fGk+B|?X zB=Kdn063=M(UaSVRq6YAE3f*rntUtyomxU_QqQJP1%!yJ?Vc)J?Rmab1Z1R{B= z9bhrxEbR~kLE9$;dX2})_a$5ywiU6M2*A{SR~Ng#%iMhOixIUsO)rExbEPuLLC*6= zfmT4DyvX=b)*1&5Jmd>LhBf}w@+N>$ii&&~@#W@Q<0T)XAb8Y`ar9)xOyJh%%&D4* z0#n+K88mzN><=}?&c=3?WZ?2NpuOqeU3lQww=A_%Qsuh01~n*|$u|AWJX-_NZbAa`X!>a!)mvTiipDW4sJw|Abi>ryb3`uD^5 z>T7Lz(A$QJ`n0f7j1qg%W3WN- zbAZr{x>@?+!-(die?*`WXX1%DpJObf?r0Te3Sc($!c@pElA7?7@E{ji3HV7!#)m7C z#>7l+5_~rE591eD^iPGE}y`^ zRvuO2{QP%!)O2_dt@T0QjnqU}^9rEP^WWOvbXAsntQ`kE8vb>Xz{zAArOk6P{$zbk zu30=YB`y5~p>Gu|d9Fv`A_xP1883~FkqMmB9w{N!7Mz_`Xxf-D$W}X<_53|jflo8) zV|&^}7Ey73rON;;ANl#~vv(no!b&z)2R|4#7uJgUvUx5$+I10OSkqa=@AZvYAaDH|hAxJ^9`j6%( z_w=ngx_JGL5{PY@Qpm<7brb}|Q$~!kF;T#lT)h6AJ)@RXDemdDv6uRRo&%TC7 z`-h$Rssz$)_>^(E_TgcUx@7zq5mm`Ue*E9e* zU@JXsDT8dkkze|up}>NnHr@+Ew4!&nk0 z5tL^@>a7e9oLj+LkUQ(FaRMUdK9E{NR4r9mJDe~O(Rp&I)~V24p@I>KeKn ze70>fX7gK_e+c$$L&mkwwcJCL5AzkfQ*eRzU`aB|`3# zQ)%_KuK?NWal2yvC9YWHe5@2T)_TzU{5!|MVDWzQ6kN)_@9sI!$ad9}W_yOi3%RF4DARl&S9z)|_!viw>EW z{)My&8>RCbYm~3yP!$?v`yZamoBiGc+RMjgto&Rhyo1XlM1NgRUx@umBm!{0s10`) z9X}Pxxy>EH&?S$Z%>^nyOSC-uCeb{$=tB$9^Wf`H&ijxHM9**6TWr%B>S6z``XST+Yqy@lhUDV?A%l%cZC$wdBSZ^mjh=ne%F!o|#<}pUTea&?Uu8(65K= z>f23DtiN%0OTo0y^6y+w+s3#-WQ3<%%bJ@v@)Rq&?Y^#`8K}}GV0~N%o8tUsK)b97 zIwPG5g8)Tpr)Ttfw!4u`oHKQf^pHP#!vgu^HE*l?2x1JC!$#5TVj}m(ttCh;&vz*m2W|JRWX^bE~_IIU3gDE)zUJP|W z3|VOYY`R*E{S6eILqI*hLK0LW@Mr_RAC*j(=0fOJBixcM`Oy6~gQ+3N0Ob!~tod)A zwQqFY*uJa(5_MRaTE&#zX&TZPBRyO0q_5qn)CnJ(7S`}I#*+BbCBg$Eu{<}L58sd5 zUybR|VqrRvT z<^q$Db!eTxy&k@#x&EP*t4w*!@5449?RI|PNb&-rk~Wzyj!Vww$7)ELGH$pSiCG5ny{E8P<3rP3TxR zxf`u02G+2gYiENo!0)dpx+;Y-Q$*+7D|+>t&}!YO>cAlV%&O}Ukr^pI|K(IH3j z;PiRj3FPC)e@ORB{V8f@QD-_oK2Q&>eOBZmo0P9mcxWVLJLoIRztFbyJ2k-_Ay?XV zX9R||-WFtI5a|=r!UF=H?Zt)7N$@4^Y$Ux( zpV>w$eMjo=%&r4asAmLOrzO(G+c;97C%{AI8!-papeRE+v%K>rw~#S_1TvLw35#yI z-?%1?mnpz54{oU2uwKitvxX8~YyIG54+G9x4?kEu!uf=N`?5*u;hX#dYO5joP&AUue;-^6WLR|5Vv5zrXZhYc&eNyNsShi*ac@ zDZYlF;LNrlns|L4SeMM}Y`HRwKSH^T#z|Y7vAwNt@7}1myw7x0Mg9~CQQTOLn?F4s zQKD#~0m=LJYx0K`IJmk_UpU&06t!uSTJp(R+Mce)kFABRfVmnNW;Ec9sVyk~#_ePm zUSi7?70pDsTpjj{>);GJK2n)KacJ)r<>fAC=k@2N#w#!ao(Z&I_O=UBG529fn0mYQ zr4oJ&c{tqkr!PAKpH{u?5u)VeJLnmk7VMYotm1H};BI5UWf$pA&9@O;XIUr zIKuyv72K2fv)X1m-;;hXo!(DlIx4c-oyCPK!+Sfx#xD1q}k~SCx@8)bHgwB37+2{D!;^>v|W&@^PAAAkxgn3 zho1Tr`kk8X-c@mVfYjvb+JfX?b3sPCujV{_o^9IRCjH3!%xDQbIK;!aVR*7I03|PI z#M};+pV#V9Ir=K8YWD}>v zJON$ymW?p*l7;FaNpK(L>X^x8!N&ah>}Ghuw+93IH<-PC@TuOw7;JSeF0aKC)Xw@kM77L;i{nT5Qj=>VE^iV0rTSYF?~-G^7EH z8(izUQ*piIr8x-ZM9 zP@*RDaC#*X?P(O&tKL1|!8|6K+5ezw;&?m$aRA;?HX#v&V|I1@(2~EG|0Z%f)ASuQ zN5!?y$sx3UYy|2`_|>Ia3IVV!zI@*16+sb6514s9Uc@cwyk$b6jF9U>ND4h~=Yk53 zdx{KAw@a-)RD>r)<>%OC?3(e&kUyb;kpcQBFp_9gC8vmrI=>*e$N-wNq~qLvy2SVs zzGD~zsVwHqmMc{HL9F7y%X&nsu!lb%u)p7mqiEamzZDaAI)A|+1_^C*uw1a_rsviQ za;>$sjaTyK-qE;p$kexORG+7=?q@vQBizlJlF1x%A5So{iXv9NOnwv+qr$41n;oZ(zlD8*QWU;zmx6;V3x7~*O z9uVHWck1nRaAuN|wYwlKr#%~n>pfjGUNY32uVNicKTPVSc?fuJY!=ybNV} z#g$2t#RLzPU3R!HRBXd++ngiB^MZ~5KbuEfIJO$Lf^0%YX1?}$GaNF@oK#`+5W-QF zU4pW3d|aco=&*5$9T&-h+p%)$Bj8ab(mqqKuHY^%_3?@kKEdNu9WnkKqnCsoz;# zA5po4#ik;AOWRC*0UK+5looIr=%J)Zf|rncZRl|dOxf}pTsrDUw|x!*y-bZDv{Vas<%WYH{$}AGfMLU-3j)Ki?v!fhsC^1=-5wh%+2QUh}p{w z&hIVyGLhKwsx&9sJH@b7U*jdtL47gv=31q0|J^64Wv9b#6c94!T(%v38||07?`mLi z4xt_9wjf`!+iy}8sERAe>J-W&dL;Mp6@)Xw&j46N=rTvA3CcJyi*<6~2YVC^9J)D8 zel*+v=B}DqjdLqvvYg?Z#Da5DS(05e=_&7-K$0x`ZEof25qH~a%NL&BJ%&}cCvP<0 zII+9$K)1cpcx1`b~3Lj5BizUQ`Eo>t|6Q?~PM zEu4a}2_qYe4J{dG{56t#L+@B&S{`oO0;YkYF)X_j{&|4fC&`dZMBQWnga*Y0HXGmJ z*aq*-C>nfQXku>s`Kw_=-TrI^jRe0-8PMXtv{Zou<7qnjD{NQUD4AYh1XGO4u!ORg zUAS=!7w(n%RwwYfTMpFro38nc=j`l(vdD}bucj=KG_kBn3v}c`m}Tm1h1%8-lNRrr zpw(vAx(JO&B#lpaUs`KQSyuDF@FNFjD#9&4#9Mi~HO+L=>PB;YtK)L48b?V`VZ3Hh zDFD$iuO-qc&xm`%Y13{X0>#yuQrG@cI%i?vczomoon?v(S8EfdW1$Np0AIx+p=Qw0 zpu&cIui0+Dmpuo!dG`{j7CAVij6L3W_LTNHoZQVUg6{>HkWVZ?SqPy+I@27xgjz$^ zs@%A%q%okrK8Q%Z<&?iDH*{gjH&`DY?REJZ2&aB*>iOdzgi2^Y7ry)VG-H<;wrO<=4 zV{>qb(s#1#2^J;BC!J9ZzCPkQGflWs*O2K{w`z7RyTlN5qWgRIdB|z#IH0SBGGFmP zQ7N2j+$TH>{*gv*Ty*pTB9+F3*ZZbtzc`-IXLvC#3}mGQNN$Nn z_(kZCopg22zBg6Zfb)9M;?THWHZTr-bo=E+F{hE|%Cj1v)>-nL^T79p2TLo(bu-$< zq>sSEx&x=s8B`!IoI4CriegOtaOA?v=Jt*n?;kClBKAesH$qDKOwT7H7>Chh>#crGJY7Vg}A(t&&}}U`knb z#|K`w2Paw)Y0=hgjdm*@_dM**l=Lh6?!@wQfjJqnOaJ?{jQPA{(+%=`Q^9KbQRHfBDbL@c;WJsR