Abstract For the satellite attitude estimation problem, the serious model errors always exist and hider the estimation performance of the Attitude Determination and Control System (ACDS), especially for a small satellite with low precision sensors. To deal with this problem, a new algorithm for the attitude estimation, referred to as the unscented predictive variable structure filter (UPVSF) is presented. This strategy is proposed based on the variable structure control concept and unscented transform (UT) sampling method. It can be implemented in real time with an ability to estimate the model errors on-line, in order to improve the state estimation precision. In addition, the model errors in this filter are not restricted only to the Gaussian noises; therefore, it has the advantages to deal with the various kinds of model errors or noises. It is anticipated that the UT sampling strategy can further enhance the robustness and accuracy of the novel UPVSF. Numerical simulations show that the proposed UPVSF is more effective and robustness in dealing with the model errors and low precision sensors compared with the traditional unscented Kalman filter (UKF).
Highlights The novel unscented predictive variable structure filter (UPVSF) is presented with sufficient theoretical proofs. The UPVSF is proposed based on the PVSF and UT sampling strategy. The UKF is employed to compare with the proposed two estimate methods. Simulation results show that the proposed two methods have better estimation performance than the UKF.
Unscented predictive variable structure filter for satellite attitude estimation with model errors when using low precision sensors
Acta Astronautica ; 127 ; 505-513
2016-06-23
9 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Residual Based Adaptive Unscented Kalman Filter for Satellite Attitude Estimation
British Library Conference Proceedings | 2012
|