Distributed Gaussian particle filter for target tracking in Wireless Sensor Networks

Abstract

In this paper, we present a distributed Gaussian particle filter based on Mahalanobis distance (DGPF-MD) for target tracking in wireless sensor networks. The proposed algorithm consists of three major steps. First, a sensor selection scheme is performed to reduce the cost of transmission among sensors with high accuracy. Second, a distributed Gaussian particle filter is adopted for each selected sensor to estimate the local statistics. Third, during weighted average fusion, the global estimate is based on the utility of the data provided by the member sensors, which is characterized as MD between the sensor and predicted target position. Compared with the centralized particle filters (CPFs), our experimental evaluations show that the DGPF-MD has more acceptable complexity, lower communication cost, and shorter tracking latency.