An Adaptive Extended Kalman Filter Based on Variational Bayesian and Sage-Husa Prediction Algorithms

An adaptive extended Kalman filter based on variational Bayesian and Sage-Husa prediction algorithms is proposed to address the problems of motion model mismatch and high-level environmental noise interference in target tracking. It combines variational Bayesian inference and Sage-Husa Kalman filtering. This algorithm can adaptively modify the present motion model and uncertain measurement variance caused by stochastic ocean environmental noise. Through simulation results, it has been verified that the tracking accuracy and robustness of this method are superior to traditional Kalman filter tracking estimation methods.