BDS快速钟差预报方法对比分析

Comparative Analysis of BDS Rapid Clock Bias Forecasting Methods

随着北斗卫星导航系统(BDS)的全面开通,研究适用于BDS卫星原子钟的钟差预报方法具有重要意义。针对BDS卫星钟差数据的特性,首先对常用的多项式模型和灰色模型进行介绍推导,然后尝试引入指数平滑法进行快速钟差预报,结合算例对比分析各模型的预报精度和特性规律,结果表明:多项式模型在BDS快速钟差预报中能够取得良好效果,平均预报精度好于其他模型;灰色模型受卫星原子钟类型影响较大,短期预报精度有明显波动;指数平滑法在短期预报中的精度略差于多项式模型,针对性能较差的钟的预报效果要好于灰色模型,在长期预报中的预报效果还有待进一步研究。

With the full operation of the beidou satellite navigation system(BDS),it is important to study the method of clock bias prediction applicable to the atomic clocks of BDS satellites.In view of the characteristics of BDS satellite clock bias data,the common polynomial model and gray model are introduced and deduced at first,and then exponential smoothing method is intro⁃duced to fast clock bias prediction.Combined with calculation examples,the prediction accuracy and characteristic law of each mod⁃el are compared and analyzed.The results show that the polynomial model can achieve good results in BDS fast clock prediction,and the average prediction accuracy is better than other models.The grey model is greatly affected by the type of satellite atomic clock and the short-term prediction accuracy fluctuates obviously.The exponential smoothing method is slightly less accurate than the polynomial model in short-term forecast,and better than the grey model in long-term forecast for clocks with poor performance.