Development of robust and energy efficient algorithms for wireless sensor networks

Abstract

newline In wireless sensor networks (WSNs), the sensor nodes estimate the parameters of newlinetheir corresponding adaptive filters to achieve the desired response for the event of newlineinterest. Distributed parameter estimation in WSNs is proposed to minimize the newlinecommunication overheads in the network and to make the algorithm energy efficient. newlineIn literature, distributed estimation of finite impulse response (FIR) systems newlinehas been studied as it is inherently stable. But in many sensor network-based applications newlinesuch as target tracking and fast rerouting, feedback or recursive systems are newlineused. Therefore the parameter estimation of feedback system in WSNs have been newlineaddressed in this thesis mainly focusing on reduction of communication overhead to newlinemake system energy efficient, and then robustness to the impulsive noise. newlineThe formulation for distributed estimation of feedback systems parameter is first newlinediscussed. Diffusion and incremental way of cooperation is incorporated for the newlineestimation of parameters of infinite impulse response (IIR) system in WSN. Subsequently newlinethe poor performance demonstrated by these conventional distributed techniques newlinein less connected sensor network known as sparse network has been dealt newlinewith multihop diffusion technique. The gradient based distributed LMS algorithms newlinefails for IIR systems particularly when the poles are near to unit circle. Because