提出一种基于多模型自适应估计~(MMAE) 的星敏感器低频误差~(LFE) 校准方法. 星敏感器低频误差主要是由空间热环境的周期性变化造成 的, 会对卫星姿态确定精度造成显著影响. 低频误差的影响可以通过扩维卡尔曼滤波~(AKF) 进行校准. 但是, 在星敏感器观测量中不存在低频误差的情况下, AKF 的姿态估计精度往往不及传统卡尔曼滤波~(KF). 针对这一问题, 将~KF 与 ~AKF 相结合, 设计了基于~MMAE 的姿态确定滤波算法, 该算法能够根据星敏感 器在轨误差特性自适应的选择~KF 或~AKF 算法进行滤波. 仿真结果表明, 所提 算法综合性能优于~KF 和~AKF, 适用于对姿态确定精度要求较高的高分辨率对地 观测卫星.

This paper presents a multiple-model adaptive estimator (MMAE) to calibrate the star sensor low frequency error (LFE). The star sensor LFE, which is caused primarily by the periodic thermal distortion, has a great impact on satellite attitude determination accuracy. The unfavorable effect of the LFE can be partly eliminated by using the calibration algorithm based on the augmented Kalman filter (AKF). However, the AKF may be worse than the traditional Kalman filter (KF) in the absence of the LFE. In order to cope with this problem, the MMAE is applied first time for combining the AKF and the KF in the spacecraft attitude determination system, such that satisfactory performance can be achieved in different operating scenarios. It is shown via numerical studies that the presented algorithm outperforms the AKF and the KF. The calibration algorithm is applicable for satellite attitude determination.