Efficient adaptive strategies over distributed networks

Abstract

Distributed wireless sensor networks finds many remote sensing applications like battle field surveillance, target localisation, environmental monitoring, precision agriculture, smart spaces and medical applications. Due to their vast range of applications efficient design and implementation become the current area of research.A distributed network consists of certain number of processing elements called nodes.These nodes are distributed over a geographical area which collects the information for particular phenomena and communicates with other nodes of the network to arrive at estimation the parameter. A network needs effective and efficient designs to function properly with the limited available resources.In this thesis,we review some of the computationally efficient adaptive distributed strategies developed using incremental partial update techniques. The schemes mentioned here solve the problem of linear estimation with less number of computations in a cooperative fashion. In a distributed network each node contains local computing equipment which estimates and shares them with other nodes. The resulting algorithms are less complex in competitions and in communication because of Incremental partial update algorithms and each node communicate with immediate node only. Computational complexity analysis is evaluated and performance characteristics of each algorithm are given with computer simulations. Simulation results show that with a small degradation in performance, a considerable amount of computational complexity is reduced.