Decentralized Information Particle Filter for Passive Tracking in Sensor Networks

A decentralized information particle filtering (IPF) algorithm for improving the passive tracking performance and reducing communication amount in wireless sensor networks is proposed in this paper. Dynamic clusters are organized according to the current position of the target, and the decentralized information extended Kalman filter is used to deal with passive tracking problem. The information filter framework is used to incorporate the newest observation into the proposal distribution of the IPF, and detailed implementation steps of the IPF are deduced based on dynamic clustering scheme. Computer simulation is conducted to compare tracking performance and analyze communication amount. Simulation results show that the dynamic clustering structure reduces the communication amount during the tracking, and the IPF has similar good performance in tracking accuracy but lower communication cost than the centralized particle filter.