编队卫星自主相对导航简化强跟踪UKF滤波算法

针对编队卫星自主相对导航过程中存在模型不确定性及噪声统计特性时变导致EKF滤波算法估计精度降低、鲁棒性较差的问题,以STF为理论框架,设计了简化强跟踪UKF(SSTUKF,simplified strong tracking unscented kalman filter)滤波算法。该算法基于STF的等价表示来计算次优渐消因子,避免了计算Jacobi矩阵,通过在线实时调整滤波增益矩阵,确保系统理论模型偏离实际模型时输出残差序列相互正交,增强算法的鲁棒性,并且对时变的噪声统计特性不敏感。结合CW系统模型线性特点,用标准卡尔曼滤波中的时间更新代替相应的UT变换过程,在提高估计精度的同时有效地降低了运算量,增强算法的实时性。仿真结果验证了该算法的有效性。

Simplify Strong Tracking UKF Algorithm for Satellite Formation Autonomous Relative Navigation

In view of model uncertainty and time-varying noise will result in the decrease of estimate accuracy and poor robustness for EKF algorithm in satellite formation autonomous relative navigation, design a simplify strong tracking UKF algorithm on the theoretical framework of STF, it uses STF equivalent representation to calculate suboptimal fading factor, avoids to compute the Jacobi matrix, by adjusting filter gain matrix online, ensures the output residual error sequence is orthogonal to each other, enhances the robustness of the algorithm, and is insensitive to the time-varying noise. Combine the linear characteristic of CW model, Unscented Transform is replaced by standard Kalman filter in time update process, it improves estimation accuracy while reduces the amount of computation effectively, and enhance the real-time performance of the algorithm. Finally the simulation results verify the effectiveness of the algorithm.