Efficient and robust solutions for sensor network detection and localization

摘要: Chin, Jren-Chit. Ph.D., Purdue University, August, 2010. Efficient and Robust Solutions for Sensor Network Detection Localization. Major Professor: David K. Y. Yau. The ability to quickly detect locate stealthy radiological dispersal devices (RDDs) allows authorities disarm remove the RDDs before they can be detonated. Traditionally, detection of was accomplished by using expensive cumbersome radiation sensors strategically located in surveillance area. However, with recent advancements wireless technologies sensing hardware, deploying a large scale sensor network small is now becoming reality. In this dissertation, we study methods sources accurately sensors. Localization single source achieved three noiseand error-free environment. When both noise errors are considered, present closed-form solution that outperforms existing algorithms. more than available, an efficient algorithm exploit additional data, order further improve robustness accuracy localization. To localize multiple network, propose hybrid formulation particle filter mean-shift technique, achieve several important features which address major challenges faced localization First, our able maintain constant number estimation (source) parameters even as K increases. Second, “learns” from estimated instead relying on statistical estimations. Third, presence obstacles