空间转移飞行器自主导航系统SINS/GPS/SS的卡尔曼滤波器降阶方案

One Redueced-Order Plan of Kalman Filter of SINS/GPS/SS for Space Transfer Vehicle Autonomous Navigation

基于Kalman滤波的Strapdown Inertial Navigation System/Global Positioning System/Star Sensor(SINS/GPS/SS)组合导航系统已被研究用于满足空间转移飞行器自主导航能力的需求。为提高该方案实时性,论文提出两步降阶简化设想。第一,在星敏感器精度较高且保持稳定的前提下,认为绝对姿态(四元数)误差较小,可忽略组合系统速度误差通道中的姿态四元数误差,实现姿态四元数误差与速度和位置误差的解耦,进而将原滤波器简化为分别估计姿态四元数误差和估计速度、位置误差的两个完全独立的降阶滤波器;第二,针对估计位置、速度误差的滤波器,采用噪声等效思想,简化加速度计误差模型,进一步降低滤波器维数。此时,原16维滤波器将被化分为两个同时运行的7维和6维降阶滤波器。仿真表明在计算量明显下降的同时,该降阶滤波器和原滤波器估计效果却相当,证实这种思想具有一定的可行性。

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 fiher model. The quaternion error state in original SINS/GPS/SS fiher 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.