大空间GPS网共模误差剔除研究

摘要

利用叠加滤波、PCA、KLE 3种方法对欧洲区域平均基线长度为2 000 km的大空间GPS网12个IGS站9 a(2006-01～2014 12)的观测数据进行处理,分析对比3种方法剔除共模误差的效果.结果显示,3种方法剔除百分比分别为25.87％、41.48％、35.72％,滤波后坐标精度明显提高;对于较大的GPS观测网,相对于叠加滤波,PCA和KLE法效果更好,可以显示不同站点的空间特征.%The data of 12 continuous GPS reference stations from Jan 1st,2006 to Dec 31st,2014 are computed and compared using stacking filtering,PCA,and KLE methods.The mean baseline of these GPS sites is 2 000 km.After common mode error is removed,the mean RMS of coordinate time series declined 69 ％,60％ and 41％ in N,E,U directions respectively,and position precision is highly improved.For large-scale GPS Networks,PCA and KLE methods can reveal spatial patterns of different sites,more effectively than the stacking filtering method.