Sujet : Decision on your Manuscript #CLUS-94 Date : 5 Feb 2010 15:47:58 -0500 De : Cluster Computing Pour : Ahmed.Mostefaoui@pu-pm.univ-fcomte.fr Copie à : hariri@ece.arizona.edu, I.U.Awan@bradford.ac.uk Dear Dr. Ahmed Mostefaoui: We have received the reports from our advisors on your manuscript, "Distributed Data Fusion in Sensor Networks : An Asynchronous Iterative Approach". With regret, I must inform you that, based on the advice received, your manuscript cannot be accepted for publication in Cluster Computing. Attached, please find the reviewer comments for your perusal. I would like to thank you very much for forwarding your manuscript to us for consideration and wish you every success in finding an alternative place of publication. Sincerely yours, Salim Hariri Cluster Computing Comments for the Author: Reviewer #1: The paper addresses the issue of distributed data fusion in wireless sensor networks. This is based on a consensus algorithm that allows nodes to converge to some average of sensors measurements of event observed. The authors claim that it is robust to asynchronous measurements and adopt to dynamic topology changes. This work is basically an extension to the author's previous work [14], which they have not explicitly mentioned. This study like the others provides fusion algorithm, which merely provides average of measurements. Therefore, these are applicable to specific applications of wireless sensor networks in which event is detected in a little bit wider range by a number of collocated sensor nodes and change in phenomenon is relatively low. For example, light or temperature monitoring are the target applications for these studies. Hence, the cope is limited. I suggest the authors to incorporate some other fusion functions like deviation or extreme values to make the fusion algorithm more realistic. Moreover, convergence to average of initial states is not the right measure of event information rather it should be converged to recent measurement. There are some grammatical mistakes and writing needs to be improved. Assumptions 1, 2 and 3 are given but it would be helpful to discuss the consequences of these assumptions on the fusion algorithm. There is no comparison with the existing study to see the performance gain of this work. The results should also be provided to show the time taken for convergence in addition to no. of iterations. This is important in case of asynchronous fusion algorithm due to unanticipated communication delay. Reviewer #2: The authors started with a good flow but in the middle and specially in the performance analysis they lost the command. There are a few spelling mistakes and reference are missing in the text. The authors also claimed that they have incorporated the dynamically changing topology, but no results have shown in this regards.The results have not analyzed and explaind properly. Results shown in fig 4 and 5 are identical with only axis changed which are not under stood at all. Convergence time is take as the the simulation time which is not correct and need to develop an accurate model. The authors are advised to co-relate their mathematical conclusions with the simulation results accurately and explicitly.