Collaborative multi-target tracking in wireless sensor networks

Abstract

A collaborative variational/Monte Carlo scheme is proposed to solve the multi-target tracking (MTT) problem in wireless sensor networks (WSNs). The prime motivation of our work is to balance the inherent trade-off between the resource consumption and the accuracy of the target tracking. For the sake of resource efficiency, we reduce the MTT problem to distributed cluster-based variational target tracking when the targets are far apart; and switch to data association only when the targets are gathered, leading to ambiguous measurements. The sequential Monte Carlo (SMC) method is employed to assign the ambiguous measurements to specific targets or clutter based on association probabilities. The associated observations are then incorporated by the variational filter, where the distribution of involved particles is approximated by a simple Gaussian distribution for each target. In addition, considering the situation that the number of targets is varying, an hypothesis test is integrated into the collaborative scheme, to deal with the cases of arrivals of new targets and disappearances of the tracked targets. The effectiveness of the proposed scheme is evaluated and compared with the classic SMC MTT algorithm in terms of tracking accuracy, computation complexity and energy consumption.