针对噪声特性未知的多传感器航天器姿态估计过程中互协方差未知的问题,提出一种鲁棒的协方差交叉(CI)融合算法.首先采用容积卡尔曼滤波(CKF)器获取局部的估计信息;然后以最小化非线性性能指标为原则求取局部的估计信息权重;最后使用CI算法融合各局部估计信息.此外,对于由四元数描述航天器姿态时存在的冗余问题,采用了以误差四元数和误差广义罗德里格参数相互切换的方法来替代.仿真结果验证了所提出算法的有效性.
This paper presents a robust covariance intersection(CI) fusion algorithm for the unknown covariance in the process of multi-sensor spacecraft attitude estimation with unknown noise characteristics. Firstly, the cubature Kalman filter(CKF) is used in the algorithm to get the local estimation information. Then, the local estimation information weights are obtained based on the principle of minimized non-linear performance. Finally, the CI algorithm is used to fuse the local estimation information. In addition, a method of switching between error quaternion and error generalized Rodrigues parameter is used to instead the quaternion which is redundancy to describe the spacecraft attitude. The simulation result shows the effectiveness of the proposed attitude fusion algorithm.