Maneuvering Target Tracking based on Adaptive Cooperative Cubature Kalman Filter

Abstract

For tracking a target with a sudden maneuver change by using multiple sensors, a distributed estimation approach based on adaptive cooperative cubature Kalman filter (ACCKF) is proposed. For alleviating tracking performance degradation caused by target abrupt maneuver, an adaptive cubature Kalman filter is designed by adding a time-varying fading factor, which can adjust filtering gain matrix in real-time. In ACCKF, a consensus-based distributed cubature Kalman filtering algorithm is developed, where a consensus weight matrix is present in the sensor network with communication topology. The ACCKF guarantees that all the local sensors can achieve a consensus in state estimation. Compared to centralized estimation approach, ACCKF can reduce the communication and computation burden. Simulation results show that ACCKF has good performance in tracking the target with a sudden maneuver change. Compared with the existing estimation approaches including the adaptive cubature Kalman filter, ACCKF has the best overall filtering performance.