Target Tracking Using a Distributed Particle-Pda Filter With Sparsity-Promoting Likelihood Consensus

Abstract

We propose a distributed particle-based probabilistic data association filter (PDAF) for target tracking in the presence of clutter and missed detections. The proposed PDAF employs a new “sparsity-promoting” likelihood consensus that uses the orthogonal matching pursuit for a sparse approximation of the local likelihood functions. Simulation results demonstrate that, compared to the conventional likelihood consensus based on least-squares approximation, large savings in intersensor communication can be obtained without compromising the tracking performance.