Radar Target Tracking Based on Some Multi-Dimensional Adaptive UKF

When the system model is inaccurate or the noise is uncertain, the traditional UKF algorithm has problems such as decreased filtering accuracy or even divergence. Aiming at these problems, this paper proposes an UKF algorithm based on multi-dimensional adaptive factors (SMA-UKF). Firstly, the sufficient conditions for establishing strong tracking UKF are expounded combined with UKF algorithm and strong tracking filtering principle. Furthermore, some multi-dimensional adaptive factors are introduced into the single-step forward prediction covariance matrix, and their calculation methods are designed respectively. Finally, the effects of SMA-UKF, strong tracking UKF algorithm (ST-UKF) and UKF algorithm in target tracking are simulated and compared under the condition that the system model and noise are inaccurate. The results show that SMA-UKF can automatically judge and adaptively adjust the process noise, and achieve good tracking of the target.