Objectives   Aiming at the unstable error of the extended Kalman filter (EKF) and the limited representation ability of a single motion model in a scenario involving complex changes of a ship's motion state, a ship trajectory tracking algorithm based on an interactive multi-model (IMM) square root cubature Kalman filter (SCKF) is proposed.

  Methods  The SCKF is introduced to replace the EKF in performing the trajectory tracking of automatic identification system (AIS) data; the constant velocity model (CVM), current statistical model (CSM) , constant turn rate model (CTM) and improved CTM are combined using an interactive multi-model framework, and three combined models are constructed to characterize the motion state of the AIS trajectory. Trajectory tracking experiments are carried out using the three combined models.

  Results  The results show that in Trajectory 6, the root mean square error (RMSE) of the position information of the SCKF is smaller than that of the EKF, and the accuracy is improved by 30.06% when Combined Model 1 is used to track the trajectory with varying heading, heading rate and velocity; and when using Combined Model 3, the SCKF has the smaller fluctuation range of RMSE compared to the position information using the EKF in Trajectory 6, and the error value is reduced by 60.80%. Combined Model 3 has the best performance, but its computation is large. In a complex trajectory experiment at constant velocity, the performance of Combined Model 2 is close to that of Combined Model 3.

  Conclusions  The proposed method can improve the accuracy of AIS data and ensure the stability of AIS data error fluctuation, making it possible to improve ship motion tracking and monitoring.