An adaptive radar surface vessel tracking method with interacting multiple model and gated memory

Radar is a widely used state measurement unit in intelligent vessel systems and maritime surveillance. However, in new radar tracking scenarios, the statistical characteristics of noise are usually unknown. As a key target tracking method, the Interacting multiple model (IMM) filter tends to experience significantly degraded performance in surface vessel tracking due to the inefficiency of model switching under noise interference. To solve the problem, this paper proposes an adaptive method that can improve the accuracy and robustness of target tracking, by correcting the noise matrix of the rank Kalman filter (RKF) and the transition probability matrix (TPM) with exploitation of historical information in real-time. Then a gated memory mechanism is introduced in the transition probability correction function (TPCF) to accelerate the model jump response by updating and resetting memory adjustments. Numerical simulation and experiment show that the proposed algorithm can effectively reduce the negative impact of unknown noise on IMM and improve target tracking accuracy. Compared with the IMM-RKF algorithm, the proposed algorithm reduces the state estimation errors in position and velocity by 15.1% and 26.3%, respectively.