GPS坐标时间序列共模误差分离方法比较及其对噪声模型建立的影响  

Comparison of Common Mode Error Separation Methods for GPS Coordinate Time Series and Their Influences on Noise Model Establishment

作　　者：舒颖 花向红[2,3] 闵阳[1] 滕焕乐[1] 吴石军[1] SHU Ying;HUA Xianghong;MIN Yang;TENG Huanle;WU Shijun(China Railway Siyuan Survey and Design Group Co.,Ltd.,Wuhan 430063,China;School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China;Hazard Monitoring&Prevention Research Center,Wuhan University,Wuhan 430079,China)

机构地区：[1]中铁第四勘察设计院集团有限公司,湖北武汉430063 [2]武汉大学测绘学院,湖北武汉430079 [3]武汉大学灾害监测和防治研究中心,湖北武汉430079

出　　处：《测绘地理信息》2023年第3期20-24,共5页Journal of Geomatics

基　　金：国家自然科学基金(41674005,41871373);中铁第四勘察设计院集团有限公司科技研究开发计划项目(2019D082);中国铁建股份有限公司重大科技专项(2019-A02);江西省自然科学基金(2017BAB203032);常温常导高速磁浮精密工程测量技术研究(2019K123)。

摘　　要：为快速有效地分离GPS坐标时间序列中的共模误差(common mode error,CME),采用区域堆栈滤波法、加权堆栈滤波法、相关加权叠加滤波法、距离加权滤波法、主成分分析(principalcomponentanalysis,PCA)5种滤波方法对GPS坐标时间序列进行CME处理;基于MATLAB设计了相应的GPS坐标时间序列CME分离工具。并以8个GPS基准站的坐标序列为对象,对去除CME后的GPS坐标序列进行噪声模型分析。结果表明,这5种方法能有效降低各站点坐标时间序列的不确定性,提高坐标序列精度,相比其他方法,PCA法滤波效果更好;此外,去除CME后的时间序列最佳噪声模型发生了改变,且GPS站坐标序列噪声模型呈现出多样性并存在个体差异。To separate the common mode error(CME)of GPS coordinate time series quickly and effectively,we use five filtering methods to process the CME of GPS coordinate time series:area stack filtering method,weighted stack filter⁃ing method,correlation weighted stack filtering method,weighted distance filtering method and principal component analysis(PCA),respectively.And we design the correspond⁃ing CME separation tool of GPS coordinate time series based on MATLAB.Taking the coordinate series of 8 GPS refer⁃ence stations as the objects,we analyze the noise models of GPS coordinate series after removing CME.The results show that the five filtering methods can effectively reduce the uncertainty of coordinate time series of each station and im⁃prove the precision of coordinate series.Compared with other methods,PCA has better filtering effect.In addition,the best noise model of the time series after removing CME has changed,and the noise model of the coordinate series of GPS station presents diversity and individual differences.

关 键 词：GPS坐标时间序列 共模误差分离 主成分分析 噪声模型 

分 类 号：P228[天文地球—大地测量学与测量工程]