Abstract:

Considering the fact that the computation resources of an on-orbit computer are limited, this paper proposes a fault detection and isolation method based on a Kalman filter to reduce the computation load of the fault diagnosis for a satellite attitude determination system. In the proposed method, a novel additive Kalman filter is designed based on the attitude kinematics, and then a sensor fault diagnosis method is presented using the simplified filter scheme by integrating the Kalman filter with the simplified observer scheme (SOS). The proposed fault diagnosis method not only can detect and isolate the gyro fault, but also can diagnose the fault of the star sensor. In this method, only one filter is used to achieve the fault detection and isolation. Moreover, it is easy for computation and hence is useful in on-orbit implementation. The proposed method is verified by an attitude determination system consisting of a tri-orthogonal-one-slant gyro component and a star sensor. The simulation results show the effectiveness of the proposed method.

Key words: Attitude determination system, Fault diagnosis, Kalman filter