Attitude Adaptive Robust Strong Tracking Kalman Filter Based on Unmanned Surface Vehicle Navigational Radar Target Tracking

The information such as the target’s position and velocity as detected by the radar carried on the unmanned surface vehicles (USVs) are sensitive to the state disturbance and the sensor noise generated from USVs attitude change. The Kalman filter (KF), as a traditional target tracking algorithm, still needs to be improved in terms of accuracy, effectiveness and stability due to its inability to adaptively adjust the noise covariance matrix. In this paper, an attitude-adaptive robust strong tracking Kalman filter (A-ARSTKF) algorithm is proposed, which can solve the problems of inaccurate radar measurements caused by USV attitude changes and improper filter gain estimation to improve the filtering accuracy of the algorithm. Firstly, an attitude improvement method that uses the turning angle information to correct the sensor noise covariance in real time is proposed. Then, a weakening factor is introduced into the A-ARSTKF algorithm for updating the fading factor to better regulate the one-step prediction noise covariance matrix. Finally, the adaptive factor is employed to adjust the Kalman gain and the state estimation error covariance matrix, which enhances the reliability of the A-ARSTKF algorithm. The simulation results show that the A-ARSTKF algorithm effectively eliminates the influence of the state disturbance and the sensor noise caused by the USV's own attitude change, and has strong effectiveness, stability and accuracy.