基于经验模式分解和最小二乘支持向量机的卫星钟差预报  

The Satellite Clock Bias Forecast Based on Empirical Mode Decomposition and Least Squares Support Vector Machines

作　　者：雷雨[1,2,3] 赵丹宁[1,3,4] 

机构地区：[1]中国科学院国家授时中心,西安710600 [2]中国科学院时间频率基准重点实验室,西安710600 [3]中国科学院大学,北京100049 [4]中国科学院精密导航定位与定时技术重点实验室,西安710600

出　　处：《天文学报》2014年第3期216-227,共12页Acta Astronomica Sinica

基　　金：国家自然科学基金项目(10573019)资助

摘　　要：针对卫星钟差(satellite clock bias,SCB)呈现非线性、非平稳变化的特性,提出结合经验模式分解(empirical mode decomposition,EMD)和最小二乘支持向量机(least squares support vector machines,LSSVM)的钟差预报方法.首先对钟差相邻历元间作一次差,并利用经验模式分解将差分序列分解成若干不同频率的平稳分量,分解后的分量突出了差分序列不同的局部特征;然后根据各个分量的变化规律,选择合适的核函数和相关参数构造不同的最小二乘支持向量机模型分别预报;最后将各分量预报值叠加得到一次差预报值,再将其还原得到钟差预报值.实验结果表明,所提方法与常用的二次多项式(quadratic polynomial,QP)模型、灰色系统(grey model,GM)模型和单一的最小二乘支持向量机模型相比,具有较高的预报精度和较强的泛化能力.In order to solve the nonlinear and non-stationary characteristics of satellite clock bias（SCB）,a hybrid model combining the empirical mode decomposition（EMD） and least squares support vector machines（LSSVM） for the SCB forecasting is proposed in this paper.The main ideas are as follows：the single difference sequence is firstly obtained by making difference between two SCB values of adjacent epoch,and then the EMD is used to decompose the difference sequence into several intrinsic mode function（IMF） components,and one residual component.Secondly,the LSSVM are constructed to forecast these IMFs and residual values individually,and then all these forecasted values are aggregated to produce the forecasted value for the single difference.Finally,the forecasted single difference sequence is recovered to the corresponding predicted SCB.The GPS satellites are taken for example,and the prediction experiments are carried out so as to verify the feasibility and validity of the proposed algorithm.The simulation results show that the proposed EMDLSSVM model can be employed to predict the SCB effectively,whose predicted accuracy is better than those of the quadratic polynomial（QP） and grey models,as well as the LSSVM model without time series decomposition.

关 键 词：天体测量学:时间 方法:数据分析 

分 类 号：P127[天文地球—天体测量]