Multi-Rate Distributed Fusion Estimation for Sensor Network-Based Target Tracking

This paper is concerned with the multi-rate distributed fusion estimation for maneuvering target tracking in wireless sensor networks (WSNs). A multi-rate fusion strategy and a hierarchical two-stage fusion structure are presented for the energy-efficiency and the tracking-accuracy consideration. In the first stage, a local modified strong tracking filtering estimator is designed to obtain a local estimate in each cluster head in the WSNs. The uncertainties in the system modeling and noise covariances are compensated by the fading factor to improve the robustness of local estimators in each cluster. In the second stage, a multi-rate fusion estimator is designed to generate a fused estimate with a higher estimation precision. An E-puck robot tracking platform is designed, and both simulations and experiments are presented to show the effectiveness of the proposed method and platform. It is shown that the fusion estimation method is able to provide satisfactory estimation precision with reduced sampling and estimation rates.