Track retrodiction for HFSWR

In this paper, we explore the possibility of track retrodiction to improve the track performance. To make use of the extra multi-frame data, we borrow the Rauch-Tung-Striebel (RTS) algorithm for the track retrodiction purpose. Using a converted extended Kalman filter (CMEKF) as the baseline, its retrodicted version, namely the retrodicted CMEKF, shows a significant improvement in performance. Analysis shows that the position root mean squared error (RMSE) is reduced by 30% and the RMSE in velocity is reduced by 25% when seven frame retrodiction is used. Beyond 10 frames, the reduced RMSE becomes negligible.