考虑星间测量的航天器自主导航并行滤波器

Parallel filters for autonomous navigation using relative measurements between satellites

中文关键词  卡尔曼滤波  自主导航  参数估计  并行计算

英文关键词  Kalman filter  autonomous navigation  parameter estimation  parallel computing

中文摘要

      针对基于星间测量的航天器自主导航问题, 本文考虑测量中存在野值的情况, 提出了一种轨道根数辅助估 计的并行无迹卡尔曼滤波算法. 系统由两个并行滤波器组成, 通过副滤波器的状态估计识别观测野值, 进而在主滤 波器中修正导航定位结果. 文章选择了星间相对观测两卫星编队的基本构型, 研究了算法的阈值参数选择, 对不同 参数条件下的滤波结果进行了对比. 数值仿真说明了该算法在观测量变化率较大时能够有效降低连续野值对自主 导航系统的影响, 和传统算法相比具有更高的滤波精度和收敛速度.

英文摘要

      In order to reduce the influence of observation outliers on autonomous navigation using relative measurements between satellites, an parallel unscented Kalman filter based on orbital elements estimation is proposed in this paper. The parallel unscented Kalman filtering system is composed of two parallel filters. Firstly, the outliers in observation are identified by the parameter estimation of sub filter. Then the system states which are position and velocity of satellites are corrected in the major filter. A two satellites formation based on relative position measurements is selected as the basic system configuration. This paper also investigates the parameter selections of the parallel filter and compares the filtering results on different parameters combination. Numerical simulations show that the algorithm can effectively reduce the influence of series outliers in the autonomous navigation system when observations change rapidly. Compared with the traditional algorithm, the parallel unscented Kalman converges in less time with high accuracy.