数据深度加权半参数估计法在方差分量估计中的应用研究  

An Application of Semi-parametric Estimator with Weighted Matrix of Data Depth in Variance Component Estimation

作　　者：潘晓刚[1,2] 王炯琦[1] 周海银[1] 

机构地区：[1]国防科学技术大学九院,长沙410073 [2]北京航天飞行控制中心动力学技术重点实验室,北京100094

出　　处：《天文学报》2013年第3期245-260,共16页Acta Astronomica Sinica

基　　金：国家自然科学基金项目(61004081);航天飞行动力学技术重点实验室开放基金资助项目(2012afd1008);湖南省教育厅科研项目(11C0787)资助

摘　　要：针对天地基联合测控中多源异质数据存在多耦合性系统模型误差和粗差的情况,提出了基于数据深度加权半参数估计的方差分量估计法.利用半参数估计法补偿不确定性模型误差,同时构建数据深度权抑制粗差的影响,使半参数估计不受粗差的影响,为方差分量估计提供合理的权矩阵,使得方差分量估计能对存在不确定性模型误差和粗差的结构定权.进行了天地基对空间目标的联合测控仿真试验.结果表明,改进的方差分量估计法能有效补偿不确定性模型误差并抑制粗差的影响,确定更加合理的权值.The variance component estimation （VCE） based on semi-parametric estima- tor with weighted matrix of data depth has been proposed, because the coupling system model error and gross error exist in the multi-source heterogeneous measurement data of space and ground combined TT＆C （Telemetry, Tracking and Command） technology. The uncertain model error has been estimated with the semi-parametric estimator model, and the outlier has been restrained with the weighted matrix of data depth. On the basis of the restriction of the model error and outlier, the VCE can be improved and used to estimate weighted matrix for the observation data with uncertain model error or outlier. Simulation experiment has been carried out under the circumstance of space and ground combined TT＆C. The results show that the new VCE based on the model error compensation can de- termine the rational weight of the multi-source heterogeneous data, and restrain the outlier data.

关 键 词：天体力学 轨道计算和定轨 方法 数值 

分 类 号：P135[天文地球—天体力学]