Target Tracking Algorithms Based On Nonlinear Filtering

Due to the extensive application on military and civil industries, the problem of target tracking has gained more and more attention by many experts and scholars all over the world. There are plentiful solutions have been proposed until recently. One of the most important parts in target tracking is the filtering algorithm. Kalman filter is a time-domain method, which can get the recursive minimum mean-square estimation in the linear system with Gaussian white noise. However, systems are generally nonlinear in practice for which Kalman filter is not suitable. Therefore, the nonlinear filtering becomes a very active area of the research in target tracking.Some principles and basic problems of target tracking are introduced in this paper and the target motion models used commonly are summarized. Based on Kalman filter, the main nonlinear filtering methods including Extended Kalman Filter, Unscented Kalman Filter and Particle Filter are studied. Simulations using scenario of coordinate turn model designed in the paper prove the validity of each nonlinear filtering algorithm. The advantages and disadvantages of the algorithms are analyzed and compared.EKF-RTS and UKF-RTS smoothing algorithms are implemented to improve the performance of filtering results. The influence factors of smoothing algorithm performance are studied and the result of over-smoothing method is given. The partitioned smoothing algorithm proposed can improve the tracking effect and meet the real-time requirement.IMM-EKF algorithm and IMM-UKF algorithm are implemented by combining EKF and UKF with the interacting multiple model algorithm to adapt to maneuvering targets. By using smoothing algorithm, IMM-EKFS algorithm and IMM-UKFS algorithm are obtained. Simulation results prove the effectiveness of these algorithms in the application of target tracking. Finally, the thesis is summarized, and some further research issues are presented.