Multiple model Kalman filter for attitude determination of precision pointing spacecraft

This paper proposes a novel multiple model estimator for the satellite attitude determination system composed of gyroscopes and star sensors. The main objective is to calibrate the low frequency error of the star sensor and improve the attitude determination accuracy. In the proposed approach, first, the specific frequencies of the low frequency error is extracted according to the frequency spectrum of the estimate of the gyroscope drift obtained from a standard Kalman filter; then, a bank of Kalman filters based on multiple models is implemented to identify the amplitudes of the error signals with the specific frequencies, such that a reference model is obtained to compensate for the effects of the low frequency error. Simulation results indicate that the proposed approach outperforms the existing calibration approach based on the state augmentation technique.