PCA空间滤波在高频GPS定位中的应用研究  

Study on the Effect of PCA Spatial Filtering on High-rate GPS Positioning

作　　者：殷海涛[1,2] 甘卫军[1] 熊永良[3] 肖根如[1] 

机构地区：[1]中国地震局地质研究所地震动力学国家重点实验室,北京市德胜门外祁家豁子100029 [2]山东省地震局,济南市文化东路20号250014 [3]西南交通大学测量工程系,成都市二环路北一段11号610031

出　　处：《武汉大学学报（信息科学版）》2011年第7期825-829,共5页Geomatics and Information Science of Wuhan University

基　　金：地震动力学国家重点实验室开放研究基金资助项目(LED2008B07);国家自然科学基金重点资助项目(41030317/D0211)

摘　　要：利用主成分分析法(PCA)对高频GPS时间序列进行空间滤波,并结合Karhunen-Loeve展开法(KLE)对其结果进行判定,可有效地提取共模误差,提高单历元定位精度。通过对四川GPS连续观测网的计算分析表明,PCA方法可较好地减弱区域共模误差,并能准确地反映共模误差的空间分布,且精度优于传统的堆栈法,这对高频GPS技术的应用和发展具有重要意义。In recent years,monitoring strong ground movement based on high-rate GPS continuous observing network has become a research focus in geo-scientific field.The spatial filtering we used is the principal component analysis（PCA） to extract the CME from the high-rate GPS position time series,and the Karhunen-Loeve expansion（KLE） is used to test the components.We process the high rate data from Sichuan GPS Continuous Observing Network,the results show that this method can eliminate the CME effectively,and the accuracy is slightly higher than the stacking.The key is that the spatial response of CME can be showed accurately by PCA,which is meaningful for applications and developments of the high rate GPS technology.

关 键 词：高频GPS PCA/KLE空间滤波 堆栈法 共模误差 

分 类 号：P228.41[天文地球—大地测量学与测量工程]