亚轨道飞行器的惯性/天文/卫星多源稳健融合导航方法

A robust fusion method for suborbital spacecraft based on SINS/CNS/GNSS multi-source integrated navigation

关键词:  卫星导航  星光折射导航  多源融合导航  故障检测  鲁棒滤波算法

Key words:  Satellite navigation  Refraction celestial navigation  Multi-source integrated navigation  Fault detection  Robust filtering algorithm

摘要:

为保证亚轨道飞行器导航系统在复杂环境下的精度和鲁棒性，针对卫星导航观测量异常粗差和大气密度误差引起的混合故障问题，提出了一种惯性/天文/卫星多源稳健融合导航方法。在惯性/卫星子系统利用双因子等价权调整滤波增益矩阵，能够削弱卫星导航相关观测时故障信息对状态估计的不良影响；在惯性/天文子系统利用惯性导航定位信息，设计了一种惯导辅助的星光折射故障诊断算法，可以有效检测和隔离大气密度误差引起的星光折射导航故障。对含有上述故障信息的场景进行了仿真，结果表明，该算法能够有效处理卫星导航观测异常和星光折射导航定位故障对导航结果的影响，具有较强的容错性能。在故障条件下，定位精度可控制在70 m左右，定速精度小于0.2 m/s，定姿精度在10″以内。

Abstract:

Aiming at the mixed fault problems caused by outliers in GNSS measurements and atmospheric density errors in spacecraft autonomous navigation, this paper proposes a robust fusion method suitable for suborbital spacecraft based on SINS/CNS/GNSS multi-source integrated navigation to improve the accuracy and robustness of the navigation system under fault conditions. The method, which uses two-factor equivalent weights to adjust the filter gain matrix in the SINS/GNSS subsystem, is able to mitigate the adverse effect of fault information on state estimation when using GNSS correlation measurements. In the SINS/CNS subsystem, this method uses SINS positioning information to design an inertial-assisted RCNS fault diagnosis algorithm, the algorithm can effectively detect and isolate the navigation fault caused by the atmospheric density error. Simulation is performed on a scenario containing the above fault information, and the experimental results show that the proposed algorithm can effectively control the effects of GNSS measurement anomalies and RCNS positioning faults on navigation results. Under fault conditions, the positioning accuracy can be controlled within 70 m, the velocity accuracy is less than 0.2 m/s, and the attitude accuracy is within 10 arc seconds.