Satellite Attitude Determination in Post-processing Based on URTS Optimal Smoother

In order to provide accuracy referential attitude of remote sensing satellite for images processing, all measurements obtained by the attitude determination system are processed offline periodically to fusion temporal information effectively. RTS (Rauch-Tung-Striebel) optimal fixed-interval smoothing algorithm based on UKF (Unscented Kalman Filter) forward filter was employed for data post-processing with considering the nonlinearity of gyro/star sensor integrated attitude determination system. The simulation results show that, compared with EKF (Extended Kalman Filter), UKF shows better performance on convergence and accuracy of attitude estimation under the condition of large initial state errors. The influence on the attitude estimation error caused by the initial state error and the sigma-points sampling distance of UKF was weakened by the RTS fixed-interval smoothing obviously, and the accuracy of the attitude estimation was improved effectively.