介绍了一种基于修正的TIKHONOV正则化方法的GPS单历元基线解算的基本原理,对连续两天同一时段的实验数据进行静态解算和单历元解算,以静态的解算结果为最或是值,计算单历元结果与最或是值之间的偏差,构成各个基线分量偏差的时间序列,用时间序列分析软件Eviews分析了各个时间序列的正态性、互相关性和自相关性,得出结论:1)偏差服从正态分布,均值都小于1 mm,说明没有明显的系统误差或者系统误差在均值中得到很好的抵消;2)标准差最大不超过±5 mm,说明单历元解算的精度很高;3)连续两天的X偏差和Y偏差分别有较强的相关性,说明连续两天之间的重复性较好,可以认为是存在一定的多路径效应;4)连续两天的Z偏差的相关性较弱,说明多路径对Z的影响较小.
The rapid GPS baseline determination principle based improved TIKHONON regularization method using GPS single epoch observations is discussed, a baseline data was observed at same session in adjacent two days is computed using two kinds algorithm, single epoch algorithm and static algorithm. The static result is recognized as most probability value, the differences between single epoch value and most probability value are calculated, which created time series of baseline vector offsets, whose distribution is tested, whose cross-correlation and autocorrelation are determined using the time series analyses software Eviews. The conclusions are de sets are distributed normally, whose means are smaller than 1 mm, which suggest that no notable ted, or system error is balanced in time series; 2)The standard deviation is smaller than ± 5 mm that single epoch result precision is good ; 3 ) The cross-correlations of X offsets series and Y offse rived that 1 ) Offsystem error exism, which suggest ts series are obvi-ous, namely those repeatability is good, which suggest there is multipath error in X or Y offset series ; 4) The crosscorrelation of Z offsets series is unconspicuous ,which suggest that the effect of multipath on Z offset is small.