顾及粗差和不合理方差—协方差阵的改进GNSS基线网平差算法

Improved GNSS baseline network adjustment approach considering outlier and incorrect variance-covariance matrix

GNSS基线网平差的函数模型和随机模型分别是基于基线向量及其方差—协方差阵构造。然而由于种种因素,经基线解算后获得的基线向量中往往含有粗差,且获得的方差—协方差阵难以精确反映基线的精度信息。含粗差的基线向量以及不合理的方差—协方差阵势必会导致函数模型和随机模型的偏差,严重影响网平差的参数估计。本文综合考虑粗差以及不合适的方差—协方差,提出一种改进网平差算法,该算法可有效削弱粗差对参数估值的污染且能够正确恢复出基线向量的精度,平差后点位坐标精度相比传统算法有明显提升。

The function model and stochastic model of Global Navigation Satellite System(GNSS) baseline network adjustment are constructed via the baseline vector as well as its variance-covariance matrix. However, due to various factors, the baseline vector calculated inevitably contains outliers, and the corresponding variance-covariance matrix cannot accurately reflect the accuracy information of the baseline vector. In summary, inappropriate observations and variance-covariance matrix will inevitably lead to the deviation of the function model and stochastic model, which will seriously affect the parameter estimation of the network adjustment. In this paper, an improved GNSS baseline network adjustment algorithm is proposed by considering the impact of the outliers and incorrect variance-covariance matrix. The proposed approach can effectively weaken the pollution of the gross error to the parameter estimation and can correctly restore the accuracy of the baseline vector. The example analysis shows that the adjusted point coordinate accuracy has been significantly improved compared with the traditional approach.