Analysis and Application of Extracting GPS Time Series Common Mode Errors Based on PCA

Abstract

Characteristics of daily position time series which dated from 2006 to 2011 in South Central of America GPS fiducial network are researched in this paper. A spatial filtering algorithm based on principal component analysis was employed to extract and remove the common mode errors from the coordinate time series. This method promoted the accuracy and reliability of the sites coordinate. Further, we discussed the differences influence in velocity field of GPS stations by test to wipe off the common-mode error. The result shows that the common mode errors can not be neglected in time series analysis, particularly when we deal with the micro deformation, the PCA method can get rid of the common mode errors effectively and improve the reliability of result.