基于改进因子图模型的航天器组合姿态确定算法

航天器姿态确定是航天器姿轨控制、在轨正常运行的关键, 针对航天器多姿态传感器存在测量噪声非高斯分布、可能出现敏感器失效和故障等问题, 提出了一种基于改进因子图模型的航天器组合姿态确定方法。通过建立因子图模型, 将地磁/星敏/陀螺测量信息作为因子节点加入因子图模型, 利用观测蒸馏法对观测数据集进行提炼以及自适应调整, 实现对航天器的姿态确定。在复杂条件下, 该方法扩展性强, 可以实现即插即用, 合理而充分的利用其他姿态测量信息, 避免了基于卡尔曼滤波算法中的复杂系统重构过程, 从而有利于多传感器融合。实验结果表明, 在进行复杂条件下的地磁/星敏/陀螺组合定姿仿真时, 该算法可行有效; 有传感器切换时, 具有较好的动态稳定性, 实现了即插即用。该方法灵活度高, 为解决复杂条件下的多传感器组合定姿提供了新思路。

Spacecraft integrated attitude determination method based on improved factor graph model

Spacecraft attitude determination is the key to spacecraft attitude and orbit control and on-orbit normal operation. In view of the problem of the non-Gaussian distribution of measurement noise of spacecraft multi-sensors, and the problem of possible sensor failures and malfunctions, the spacecraft integrated attitude determination method based on improved factor graph model is proposed. By establishing a factor graph model, the geomagnetism/star sensitivity/gyro measurement information is added to the factor graph model as a factor node, and the observation data set is refined and adaptively adjusted by the observation distillation method for realizing the attitude determination of the spacecraft. Under the complex conditions, this method has strong scalability, can realize plug-and-play, reasonably and fully utilize other attitude measurement information, and avoid the complex system reconstruction process based on Kalman filter algorithm, which is beneficial to multi-sensor fusion. The experimental results show that the algorithm can improve the accuracy of attitude determination using magnetometer/star sensor/gyro, which has better dynamic stability, and realizes plug and play. This method is highly flexible and provides a new idea for solving the multi-sensor fusion for spacecraft attitude determination under complex conditions.