One redueced-order plan of kalman filter of SINS/GPS/SS for space transfer vehicle autonomous navigation

To meet the requirement for autonomous navigation capability of Space Transfer Vehicle, the Integrated System of Strapdown Inertial Navigation System/Global Positioning System/Star Sensor (SINS/GPS/SS), based on the Kalman Filter, has been researched. To improve the real-time running performance, the plan of two-steps reduced-order Klaman Filter of SINS/GPS/SS is put forward. First, with the hypothesis that SS precision is high and stable, the quaternion error state would be small and could be neglected in the velocity error loop of SINS/GPS/SS filter model. The quaternion error state in original SINS/GPS/SS filter would be decoupling with the state of velocity and position error. The original filter could be separated into two reduced-order filters which are independent with each other and are special for quaternion error and velocity and position errors, respectively. Second, with the equivalent noise method, the model of accelerometer would be predigested. The reduced-order filter for velocity and position errors could still decrease the state number. With the two steps, the dimension is decreased from 16 of original SINS/GPS/SS filter to 7 and 6 of two reduced-order filters. The simulation results show that filter performance of the reduced-order filters is close to the original SINS/GPS/SS filter's while the calculation is decreased obviously. The reduced-order plan is verified to be feasible.