A fault tolerant filter applied on magnetometer-based autonomous satellite navigation

Magnetic field vector is a function of the satellite's position. Autonomous LEO satellite navigating can be achieved based on the measurement data of the onboard three-axis magnetometer (TAM). For the orbit estimation based on the measurement of the geomagnetic field vector, when measurement data influenced by outliers, filtering advents deviation, convergence becomes slow and even divergent. Based on the residual orthogonal of EKF (expanding Kalman filter), fault tolerant filtering method is given. The instant modifying and fault diagnosis are implemented and the filter obtains stronger robustness. Simulation results show the effectiveness of the method.