Analyses and Comparisons of Some Nonlinear Kalman Filters in POS for Airborne SAR Motion Compensation

Position and orientation system (POS) has been used widely for airborne synthetic aperture radar (SAR) motion compensation. In order to improve the navigating precision of POS, first the nonlinear mathematics model based on quaternion error equations is established, then one method of combining unscented Kalman filter (UKF) and extended Kalman filter (EKF) and the other method of combining model predictive filter (MPF) and extended Kalman filter (EKF) are used in POS, and comparing these two methods with EKF and KF in precision and real time. The results of flight tests show that the method of combining UPF and EKF has the best precision and convergence in these four methods, and the method of combining MPF and EKF decreasing the time of computing by reducing the number of states.