Power-efficient analog forwarding transmission in an inhomogeneous Gaussian sensor network

Abstract

In a wireless sensor network (WSN), sensor power is usually limited and must be carefully managed for each intended application. In this paper, we consider the optimal power scheduling for the joint estimation of a Gaussian source by an inhomogeneous Gaussian sensor network. The goal is to minimize the total power consumption while satisfying a certain mean squared distortion constraint. We assume that sensors transmit their observations in an analog fashion: each sensor simply amplifies and forwards its noise-corrupted analog observation through an additive white Gaussian noise (AWGN) channel to the fusion center (FC), while the latter combines the received sensor messages to generate the final estimate. Such analog forwarding strategy can be shown to be optimal in the single sensor case by Shannon's separation principle. For the multiple sensor case, we derive the optimal power scheduling using convex optimization and show that it admits a simple distributed implementation. Simulations show that the proposed power scheduling improves the mean squared error (MSE) performance by a large margin when compared to that achieved by the uniform power scheduling.