首先分析了异常观测信息对参数结果的影响;然后以假设检验为手段通过残差向量构造粗差判别统计量,以一定的置信水平判定哪个观测值可能含有异常,有针对性地在相应的观测方程中引入新的平差参数。通过附加参数的平差补偿模型求出异常大小,从而控制异常数据对结果的影响。实测GPS算例表明,通过该方法能够有效地抵制参数先验信息异常的影响,并能解出异常误差的大小。
After analyzing the influence of the outliers on parameters in the GPS observations, we conducted an adjustment compensation method in detail, that is, the extra-parameters were added in the observation equation when the outliers in the GPS observations were found by means of the hypothesis testing based on the standardized residual. A GPS case was given to show that the compensation method could control efficiently the influence of the outliers in the GPS observations, and could calculate the outliers' value.