卫星出境数小时后的入境钟差预报参数,是基于24小时历史数据拟合的二次项系数。在传统导航钟参数拟合算法的基础上,根据卫星钟的频偏和频漂参数在不同拟合时段的变化特性,分别选取最佳拟合时段,提出长短期数据结合的两步法拟合算法。基于实测数据的拟合预报实验表明,两步法的预报效果优于直接法。
Quadratic model is fitted with the nearest history data of 24 hours when the satellite is outofview, and is used to predict clock offset as soon as the satellite is in view. Based on the conventional fit algorithm, the best fit arcs are selected independently for the two fitting parameters of frequency bias and frequency drift according to their respective characteristics at different length of fitting arcs. A twostep fit is proposed using both longand shortterm data sets. The fit and prediction comparisons using the real clock offset data show that the prediction performance of twostep fit is better overall than the conventional method.