On nonlinear model correction for sampex

This paper demonstrates attitude determination of the Solar, Anomalous, Magnetospheric Particle Explorer (SAMPEX) by a Minimum Model Error (MME)-based dynamic model correction method. The MME method is applied two times to obtain the system model correction terms. In the first stage, the angular velocity rates and their corrections are determined by combinations of error quaternion multiplication and Riccati transformation methods. The determined rates are not exponentially decaying, as previously assumed. Next, the estimated rates are used to determine the error in the torques acting on the satellite. The results of the MME algorithms suggest that the model error is strongly dependent on the spacecraft’s position in space. Finally, a Least-Squares Correlation (LSC) model identification method is used to correlate the model error with the spacecraft’s position parameters. The functional forms produced are then used to correct the system model. The “corrected” system model can accurately predict the attitudes of SAMPEX during periods of total sensor failure.