Abstract:

The change of target motion will lead to a low precision or divergence of the tracking algorithm. In order to improve the performance of the maneuvering target tracking, three approaches are proposed as follows. Firstly, aiming at the increase of the model error resulted by the fixed maximum acceleration in the current statistical (CS) model, an adaptive CS model is proposed. And then, on the basis of the adaptive CS model and the interacting multiple model (IMM), a new interacting multiple adaptive model (IMAM) is proposed. Adopting  two adaptive CS models, the IMAM can effectively get rid of the rapid increase of the model error caused by the abrupt change of target motion and improve the accuracy and adaptability of the model. Secondly, based on the IMAM and amendatory Kalman filter (AKF), an IMAM-AKF algorithm is proposed. By amending the estimate of state fusion in the IMAM, the IMAM-AKF greatly decreases the model error and improves the performance of the maneuvering target tracking. Finally, combining the adaptive fading Kalman filter (AFKF), the IMAM-AFAKF algorithm is proposed. The simulation results indicate that the IMAM-AFAKF has a better performance both in high and weak maneuvers.