噪声辅助EMD方法提取多路径误差模型分析  

Modeling multipath effect with a noise assisted Empirical Mode Decomposition

作　　者：薛志宏[1,2] 周蓉[3] 李广云[1] 许云燕 

机构地区：[1]信息工程大学测绘学院 [2]航天恒星科技有限公司(503所) [3]73603部队 [4]郑州测绘学校

出　　处：《测绘科学》2013年第4期55-58,共4页Science of Surveying and Mapping

基　　金：中国航天科技集团公司卫星应用研究院开放基金

摘　　要：经验模态分解(EMD)方法处理间断性信号时存在模态混叠问题,使分解结果失去物理意义。本文从理论分析和实测数据的EMD分解结果两个方面,论述了多路径误差中不可避免的存在间断性的分量。针对该问题,介绍了集合平均经验模态分解(EEMD)方法借助噪声来减弱模态混叠的思路与流程,并连续两天实测GPS数据的单历元定位坐标序列,分别用EEMD和EMD方法消除随机噪声,提取多路径重复性改正模型,结果表明,EEMD方法的效果优于EMD方法。Empirical Mode Decomposition （EMD） has been proved to be an adaptive data analysis method and be versatile in a broad range of applications. While one of the major drawbacks is the mode mixing for intermittency signals. The reason of mode mixing was expatiated upon and an alternative method named Ensemble EMD was introduced in the paper. Then the validity in ameliorating the difficulty was demonstrated through simulation data. Multipath effect of GPS observable was complex and intermittent, then the EMD and EEMD were used respectively to extract multipath model from the coordinates serial measured by GPS in two sequential days. The results demonstrated the advantage of EEMD.

关 键 词：多路径误差 恒星滤波 经验模态分解 集合平均经验模态分解 

分 类 号：P207.1[天文地球—测绘科学与技术] P228