An improved center difference Kalman filtering method for navigation and positioning of unmanned surface vessels

In the autonomous navigation system of Unmanned Surface Vessels (USVs), the positioning accuracy of traditional filtering methods significantly deteriorates when the Global Navigation Satellite System (GNSS) signal is abruptly lost or diminished due to external environmental interference, such as the presence of obscuring structures like bridges. To address these challenges, this paper proposes an improved H-infinity central differential Kalman filter algorithm with strong tracking based on the Mahalanobis distance (MD-STHCDKF). On one hand, the improved MD-STHCDKF is guaranteed to have good robustness under the worst perturbation by overcoming the limitation that the central difference Kalman filter (CDKF), which assumes noise to be Gaussian white noise, by employing the H infinity filtering theory based on the game theory design; on the other hand, a strong tracking filter based on the Mahalanobis distance is introduced to adjust the prediction error covariance and Kalman gain matrix in real-time, enhancing localization accuracy under dynamic model uncertainty and avoiding false alarms. Experimental results demonstrate that the proposed algorithm significantly outperforms the Extended Kalman Filter (EKF), CDKF, and H-infinity Central Difference Kalman Filter (HCDKF) in terms of reliability, accuracy, and robustness of navigation filtering, particularly in signal-occlusion environments.