Event-Triggered Distributed Variational Particle Filter over Wireless Sensor Networks

This paper considers the problem of estimating a nonlinearly evolutive target state based on fully distributed nonlinear wireless sensor networks. Such a problem can be naturally solved by the distributed particle filter which, however, suffers from performance degradation when model parameters are not precisely known. In this paper, the variational Bayesian inference is exploited for joint target state and model parameters estimation, results in the distributed variational particle filter. The event-triggered strategy is also proposed to substantially reduce the communication burden among sensor nodes and, at the same time, keeps the estimation performance. Simulation results verify the effectiveness of proposed method.