基于GPS测量数据的卫星在轨轨道预报算法研究

Research of a On-Board Orbit Prediction Method Based on GPS Data

为提高微小卫星的在轨轨道预报能力,针对常用的低轨近圆卫星轨道,根据解析的轨道动力学模型,基于无奇点变量的拟平均要素法,用Kalman滤波技术给出了一种卫星解析星历参数在轨估计算法,用GPS测量信息对相关星历模型参数进行在轨估计。给出了算法流程。先由外部标志判断滤波器初始化状态,若需初始化,则可基于GPS测量数据,或地面上注星历参数,或上次滤波所得星历参数进行;若初始化已完成,则对星历模型参数进行Kalman滤波,得到更新的星历参数。给出了滤波算法中轨道预报、残差计算、量测计算和UD分解的计算模型。仿真结果表明:对轨道高度450km以上的近地圆轨道,7d内的预报精度优于20km。算法具自启动(自初始化)、收敛性佳、对测量数据的采样要求不严格等优点,实用性好。

To improve the capability of micro-satellite orbit prediction, an on-board model of ephemeris parameter was proposed by Kalman filter based on analytic orbit dynamic model and quasi-mean element method without singularity for near circular low earth orbit （LEO） , which could estimate the relative ephemeris parameter using GPS data on-orbit. The algorithm flowchart was given. Firstly, the initial state of the filter was judged using external mark. If the initialization was needed？ it could be implemented by GPS data,ephermeris parameters upload by the ground or ephermeris parameters obtained in the last filtering. If the initialization had been finished, the ephermeris parameters were treated to gain the new ephermeris parameters by Kalman filtering. The computation modes of orbit predication, residue error calculation, measurement calculation and Bierman-UD decomposing were presented. The simulation results showed that prediction accuracy was better than 20 km in the 7-days prediction for the near circular LEO which the altitude was higher than 450 km. The algorithm proposed had advantages such as self-start （self initialization） , good convergence and not ridge requirement of measurement data sampling, which was practicable in engineering.