Robust Extended Kalman filter for ballistic object tracking during re-entry

Abstract

In this paper a robust Extended Kalman filter (EKF) for state estimation of a ballistic object with nonlinear dynamics is proposed. Outliers in the measurement can seriously degrade the estimation performance of conventional nonlinear filters. The proposed robust filter resists the effect of outliers to provide improved estimation. The square of Mahalanobis distance of innovation vector is taken as the judging index for robustness. Then the proposed robust algorithm has been evaluated with a benchmark problem of ballistic object tracking during re-entry phase. With the help of simulation, it is shown that when measurement outliers exist, the robust algorithm outperforms its conventional versions.