基于鲁棒双参数指数平滑法的BDS卫星钟差预报  

Prediction of BDS satellite clock bias based on robust two-parameter exponential smoothing method

作　　者：李方能[1] 梁益丰 许江宁[1] 吴苗[1] 朱兵 LI Fangneng;LIANG Yifeng;XU Jiangning;WU Miao;ZHU Bing(College of Electrical Engineering,Naval University of Engineering,Wuhan 430033,China;Beijing Institute of Tracking and Telecommunication Technology,Beijing 100094,China)

机构地区：[1]海军工程大学电气工程学院,武汉430033 [2]北京跟踪与通信技术研究所,北京100094

出　　处：《中国惯性技术学报》2024年第7期645-653,共9页Journal of Chinese Inertial Technology

基　　金：国家自然科学基金(41804076,42004067)。

摘　　要：为提高北斗(BDS)卫星钟差预报精度与稳定性,基于BDS卫星钟工作原理与数据特征,提出了一种基于鲁棒双参数指数平滑法的BDS卫星钟差预报方法。将双参数指数平滑法引入BDS钟差预报,并通过鲁棒化建模与优化初始值选取方式提升了双参数指数平滑法的抗差与适应能力,减小了数据异常干扰对钟差预报性能的影响。与常用的多项式模型、谱分析模型、灰色预测模型及超快速钟差预报产品进行对比分析,结果表明:所提方法适用于不同BDS轨道与类型的卫星钟差预报,取得了良好的预报精度与稳定性,其对BDS卫星钟的6h平均预报精度分别提升了40.3%、31.7%、63.6%和36.6%,预报稳定性分别提升了43.8%、38.9%、65.4%和33.1%。In order to improve the accuracy and stability of Beidou Satellite system(BDS)clock bias prediction,a robust two-parameter exponential smoothing method is proposed for predicting BDS satellite clock bias based on the working principle and data characteristics of BDS satellite clocks.The two-parameter exponential smoothing method is introduced into BDS clock bias prediction,and the robustness and adaptability are improved through robust modeling and optimization of initial value selection,which reduces the impact of abnormal data interference on the clock bias prediction performance.Compared with common polynomial model,spectral analysis model,grey prediction model,and ultra-fast clock deviation prediction products,the results show that the proposed method is suitable for satellite clock bias prediction in different BDS orbits and types,and achieves good prediction accuracy and stability.The 6 h average prediction accuracy for BDS satellite clocks has increased by 40.3%,31.7%,63.6%and 36.6%,respectively.The forecast stability has been improved by 43.8%,38.9%,65.4%and 33.1%,respectively.

关 键 词：北斗卫星导航系统 钟差预报 鲁棒双参数指数平滑法 钟差产品 

分 类 号：TN976[电子电信—信号与信息处理]