High Maneuvering Target Tracking Based on Augmented State Method using Adaptive Unscented Kalman Filter

Many high maneuvering target tracking approaches, such as augmented state method, are based on target motion and radar modeling in the Cartesian coordinates., The projected radar observation errors along the Cartesian coordinates are continuously increasing in the target tracking problem in noisy operation environments especially in the high maneuvering situations that the target is moving away from radar location. Whereas, in various simulations that are performed in many papers, these observation errors are considered to have constant covariance values along the Cartesian coordinates. The real radar observation vector including target range and bearing is generally stated in Polar coordinates, and the nonlinearity of the radar observations in the Polar coordinates makes it necessary to implement the nonlinear estimation approaches such as unscented Kalman filter (UKF) or extended Kalman filter (EKF). In this paper, high maneuvering target tracking is performed using the augmented state idea in the Polar coordinates based on UKF. The new proposed method also overcomes the late convergence of estimation and prevents filter divergence using the adaptive covariance matrix approach. The simulation results obtained by the proposed method for both medium and high maneuvering scenarios show more than 90  percent improvement compared with the UKF and EKF algorithms.