Faculty of Engineering and Natural Sciences
Evaluation of Local Decision Thresholds for Distributed Detection in Wireless Sensor Networks using Multiobjective Optimization
For a distributed detection in a wireless sensor network, sensors arrive at decisions about the event of interest and send their decisions to the central fusion center. The fusion center combines the incoming sensor decisions and reaches a final decision about the absence or presence of the event. For binary sensor decisions, determination of the local sensor decision thresholds is crucial. In this paper, we evaluate the set of local sensor thresholds through multi-objective optimization where the probability of error and the total energy consumption of the network are optimized simultaneously. The optimal threshold sets are generated by using a mathematical programming Normal Boundary Intersection (NBI) method and a multi-objective evolutionary algorithm Non Dominating Sorting Genetic Algorithm (NSGA-II). Simulation results show that both NBI and NSGA-II successfully obtain a set of solutions reflecting the tradeoffs between the objectives.
Engin Masazade got his BS degree from Electronics and Communications Engineering Istanbul Technical University 2003. He then obtained his MS degree from
University, Electronics Engineering Program in 2006 where he is now a PhD candidate. During his MS and Phd studies, he has been affiliated with Communication Theory and Technologies (CTT) Group whose members are supervised by Associate Professor Mehmet Keskinoz. In 2008, he received a research abroad grant from TUBITAK to collaborate with Professor Pramod Varshney and his Sensor Fusion Group at
USA to work on the “Distributed Data Fusion for Wireless Sensor Networks” project for one year. His research interests include bit-error rate estimation and cross-layer design for Multiband OFDM systems, distributed detection, estimation, localization and tracking for wireless sensor networks.
January 20th, 2008, 13:40,