顾及卫星钟随机特性的抗差最小二乘配置钟差预报算法

发布时间：2018-06-03 00:49

本文选题：卫星钟差预报 + 随机变化特性　；参考：《测绘学报》2016年06期

【摘要】：为了更好地反映钟差特性并提高其预报精度,采用抗差最小二乘配置方法建立一种能够同时考虑星载原子钟物理特性、钟差周期性变化与随机性变化特点的钟差预报模型。首先使用附有周期项的二次多项式模型进行拟合提取卫星钟差的趋势项与周期项,然后针对剩余的随机项及其可能存在的粗差,采用抗差最小二乘配置的原理进行建模,其中最小二乘配置的协方差函数通过对比协方差拟合的方法并结合试验进行确定。使用IGS精密钟差数据进行预报试验,将本文方法与二次多项式模型、灰色模型进行对比,预报精度分别提高了0.457 ns和0.948 ns,而预报稳定性则分别提高了0.445 ns和1.233 ns,证明了本文方法能够更好地预报卫星钟差,同时说明本文的协方差函数确定方法的有效性。
[Abstract]:In order to better reflect the characteristics of clock difference and improve its prediction accuracy, the robust least square collocation method is used to set up a prediction model of clock difference which can take into account the physical characteristics of space-borne atomic clock, the periodic variation of clock difference and the randomness of the clock at the same time. First, the quadratic polynomial model with periodic term is used to fit and extract the trend term and period term of satellite clock error, and then the principle of robust least square collocation is used to model the remaining random term and its possible gross error. The covariance function of the least square collocation is determined by comparing the covariance fitting method with the experiment. The prediction experiment is carried out by using IGS precise clock error data. The method is compared with the quadratic polynomial model and the grey model. The prediction accuracy is improved by 0.457 ns and 0.948 ns, while the prediction stability is improved by 0.445 ns and 1.233 ns, respectively. It is proved that the proposed method can better predict the satellite clock error, and the validity of the covariance function method in this paper is proved.
【作者单位】：信息工程大学地理空间信息学院;地理信息工程国家重点实验室;
【基金】：国家自然科学基金(41274015;U1431115) 国家863计划(2013AA122501) 地理信息工程国家重点实验室开放研究基金(SKLGIE2015-M-1-6)~~
【分类号】：P228.4

【相似文献】