Channel estimation and multiple target tracking in wireless sensor networks based on quantised proximity sensors

Abstract

This study presents a hybrid quantised variational/sequential Monte Carlo (SMC) method for multiple target tracking in quantised sensor networks, by considering channel estimation problem. SMC scheme is employed to attribute ambiguous observations to specific targets based on association probabilities. The associated measurements are then incorporated by the quantised variational filter (QVF), where the distribution of involved particles is approximated by a Gaussian distribution for each target. In the current work, the authors propose to jointly estimate the multiple target positions, the channel attenuation between one sensor and the cluster head, and optimise the number of quantisation bits used by the candidate sensor to quantise its measurement. The multiple target positions are estimated by using the hybrid quantised variational filtering/sequential Monte Carlo-based approach to data association. The channel attenuation is estimated by maximising the a posterior distribution and the quantisation is optimised by maximising the Fisher information. The computation of these criteria is based on the target position predictive distribution provided by the QVF algorithm. Numerical examples show that the quantisation combined with channel estimation improve the estimation performances and minimise the error estimation.