Psg-Dpf: Distributed Particle Filter Using Pairwise Selective Gossiping For Wireless Sensor Network

Distributed particle filter using gossiping algorithm is a robust and efficient tool for decentralized state estimation in wireless sensor networks. Reducing communication overhead without sacrificing estimation accuracy is a major challenge. In this paper, we propose a distributed particle filter by using a pairwise selective gossiping algorithm, named PSG-DPF, which updates particles according to coefficient weights calculation instead of local weights and selects the significant particles to share among the nodes. PSG-DPF guarantees that every sensor node converges to an optimal consensus estimation which can achieve to high estimation accuracy. The communication overhead is reduced by transmitting only significant particles and controlling communication iterations. The simulation results illustrate that the accuracy of our scheme approaches to the centralized SIR particle filter.