基于记忆衰减滤波的X射线脉冲星自主导航

AUTONOMOUS NAVIGATION OF PROBE BY X-RAY PULSARS BASED ON FADING MEMORY FILTERING

针对扩展Kalman滤波(Extend Kalman Filter,EKF)在处理X射线脉冲星的自主导航定轨问题上易发散的问题,本文提出了基于记忆衰减滤波的抑制发散自主导航算法。在研究了自主导航滤波中误差变化的特征的基础上,从原理上分析了预报误差随迭代次数增加而增大的原因。针对EKF存在的这一缺陷,本文应用记忆衰减滤波控制误差发散,推迟了误差发散的时间并减小了发散的幅值,使得滤波结果稳定于更精确的结果。本文的结论均通过数值计算的方法验证。

In view of the fact that the Extend Kalman Filter(EKF)is prone to be divergent in the autonomous navigation by X-ray pulsars,this paper proposes an algorithm utilizing the fading memoryfilter.Based on the study of the error pattern of the EKF with the increase of iterations,some possible reasons for the divergency are analyzed.The fading memory filter can be used to reduce the peak of error,while delaying the divergence.