Distributed Estimation in Homogenous Poisson Wireless Sensor Networks

Abstract

In this paper we develop a homogenous Poisson point process based framework to characterize the performance of distributed estimation in wireless sensor networks. Such stochastic geometry based approaches have been shown to be robust compared to idealistic grid/ cluster models in wireless networks. We consider a fading wireless channel between the sensors and the Fusion Center (FC) and characterize the distributed estimation outage in a variety of scenarios. Towards this end we consider optimal maximal ratio combining (MRC) based coherent sensor transmission and derive closed form expressions for the outage probability. Further, we extend it to scenarios with fading channel phase based partial Channel State Information (CSI) and subsequently incorporate the effects of various path-loss models in large scale wireless sensor networks to derive bounds for the outage probability. Simulation results are presented to validate the performance of the proposed analysis.