卫星自主轨道确定的自校准滤波

A Self-Calibration Filter for Satellites Autonomous Orbit Determination Systems

研究在导航敏感器测量带有系统偏差情况下的卫星自主轨道确定问题,已有的大量研究表明这类系统偏差是影响自主轨道确定系统性能的关键因素。针对这一问题,本文基于线性时变系统的可观性理论证明了系统偏差为未知常值时,自主导航系统的状态和常值偏差都是可观的。在此基础上设计了将系统偏差作为状态向量新增分量的自校准滤波器以估计轨道参数,同时校准系统偏差。数学仿真结果表明,采用自校准滤波可以将自主导航的位置精度从10km(3σ)量级提高到200-500m(3σ)。

This paper is concerned with the autonomous orbit determination for satellites with systematic errors in the measurements of navigation sensors. A great deal of studies showed that the performance of the autonomous orbit determination system is limited by these systematic errors. By treating these errors as constant but unknown observational bias, it is proved that orbital parameters and constant bias are all observable by using standard results of linear time-variant systems theory. Based upon the observability analysis result, a self-calibration filter is designed to estimate orbital parameters and correct systematic errors simultaneously. Simulation results show that the 3- δ accuracies for position can be improved to the order of 200 - 500m from the order of lOkm using the filter presented in this paper.