مقاله سال ۲۰۱۶

Abstract

We consider the problem of distributed soft decision fusion in a bandwidth-constrained spatially uncorrelated wireless sensor network (WSN). The WSN is tasked with the detection of an intruder transmitting an unknown signal over a fading channel. Existing distributed consensus-based fusion rules algorithms only ensure equal combining of local data and in the case of bandwidth-constrained WSNs, we show that their performance is poor and does not converge across the sensor nodes (SNs). Motivated by this fact, we propose a two-step distributed quantized fusion rule algorithm where in the first step the SNs collaborate with their neighbors through error-free, orthogonal channels (the SNs exchange quantized information matched to the channel capacity of each link). In the second step, local 1-bit decisions generated in the first step are shared among neighbors to yield a consensus. A binary hypothesis testing is performed at any arbitrary SN to optimally declare the global decision. Simulations show that our proposed quantized two-step distributed detection algorithm approaches the performance of the unquantized centralized (with a fusion center) detector and its power consumption is shown to be 50% less than the existing (unquantized) conventional algorithm.

I. INTRODUCTION

WIRELESS sensor networks (WSNs) are spatially deployed over a field to monitor certain physical or environmental phenomena, to detect relevant quantities and to perform decision making functions. Because of their relatively low cost and robustness to sensor node (SN) failures they are receiving significant attention. However, there are a number of different strategies as to how the test statistics from each SN will be used in order to arrive at a final decision. We will first give a brief review before introducing our proposed approach.

First consider the centralized solution where noisy observations collected from spatially distributed local SNs are sent (inter-sensor collaboration is not considered) to a global fusion center (FC) for a final decision [1]-[12]. Then, there are some recent publications [13]-[14] (in the context of estimation) that considered the effect of inter−sensor collaboration on the estimation performance. Here, the local SNs collaborate Part of the results in this paper have been presented at IET International Conference on Intelligent Signal Processing (ISP), London, UK, December 1-2, 2015 [32].

E. Nurellari and D. McLernon are with the School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, Leeds, U.K. (e-mail: elen@leeds.ac.uk; d.c.mclernon@leeds.ac.uk). M. Ghogho is with the School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, Leeds, U.K., and also with the International University of Rabat, Rabat, Morocco (e-mail: m.ghogho@leeds.ac.uk).