Reverse Prediction Adaptive Kalman Filtering Algorithm for Maneuvering Target Tracking
Kalman filter algorithm is based on hypothetical system mathematical model and noise statistical characteristics, and it easily leads to tracking error or divergence when the hypothetical and actual models do not match. In this paper, the authors propose reverse prediction adaptive Kalman filtering algorithm for maneuvering target tracking. By comparing the norm residual square of original state and that of reverse prediction state, they modify estimation state online by noise model adjustment when the norm residual square ratio is bigger than the predetermined threshold. Maneuvering target tracking simulation results show that the proposed algorithm can improve filtering precision and robustness when target maneuvering or noise increase.