Decentralized Target Tracking Based on a Weighted Extended Kalman Filter for Wireless Sensor Networks

This paper presents a weighted extended Kalman filter (WEKF) for target tracking in wireless sensor networks, where the location estimation is formulated as a weighted least squares (WLS) problem by taking weights of the local estimates based on the reliability of distance estimation and the WLS problem is solved in an iterative, decentralized manner based on the WEKF. We adopt a message passing (MP) algorithm for inter-sensor- node communication and for adaptively selecting the participating sensor nodes as the target moves around the area. During each iteration, a participating sensor node computes a target´s location estimate and passes it on to the next participating sensor node for processing to generate a new location estimate. The update process is circulated among the participating sensor nodes in the close vicinity of the target. To show the convergence behavior of the WEKF-based method, a convergence analysis is given. Computer simulation results demonstrate that the proposed scheme has better location accuracy and tracking performance than previous related methods.