An Improved Adaptive Extended Kalman Filter Used for Target Tracking
Zixuan Long, Xiaoli Zhang, Xiafu Peng, Gongliu Yang · 2019
Traditional EKF filter depends on the condition that the noise parameters are known accurately in advance or maintain unchanged during the filtering process. When the covariance of noises cannot be obtained accurately or if the noises are time-varying signals, the performance of standard EKF may be poor and even diverge. In order to solve this problem, an improved adaptive EKF (AEKF) based on fading weight factors and prior guess of limited window length is proposed to update the covariance of measurement noise in real time. Moreover, the introduction of chi-square test based on the innovation sequences makes the time of adaptive introduction more reasonable, and avoids the deterioration or even divergence of the filter. The simulation results have showed that the performance of the improved AEKF is better than that of the traditional EKF and AEKF in target tracking.