Robust-extended Kalman filter for small satellite attitude estimation in the presence of measurement uncertainties and faults

摘要

Attitude estimation is one of the key technologies for satellite navigation, and a traditional Kalman filter (KF) can spit out good estimation results under normal conditions. However, uncertainties or measurement faults may occur and affect estimation accuracy, especially for the small satellite with few sensors. In this article, a robust-extended Kalman filter (REKF) is developed for the attitude estimation of small satellite equipped with only three-axis magnetometer. By adjusting the Kalman gain via single scale factor (SSF), the newly developed filter can compensate the uncertainties or faults. In order to correct different state variables by different factors, a REKF with multiple scale factors (MSFs) is also designed, which uses a diagonal matrix instead of the SSF. Both filters are proved uniformly stochastically stable and applied for small satellite attitude estimation. Three different scenarios are simulated, and the proposed REKF algorithms are approved to be more effective and robust to deal with parameter uncertainties and instantaneous measurement faults.