GNSS高程拟合中的稳健估计粗差剔除方法

Method of gross error elimination for robust estimation in GNSS height fitting

针对GNSS高程拟合中最小二乘估计不具备抗粗差的能力,至使含有粗差数据拟合结果失真的问题,将稳健估计引入到数据预处理中,建立了稳健估计选权迭代方法。研究结果表明:该方法有效地解决了在精度允许的范围内数学模型失真这一问题,并结合具体工程实例分析证明了该方法的有效性。

In GNSS elevation fitting,the least-squares estimation does not have the capacity to resist gross errors.Therefore the mathematical model might be distorted,and the results of fitting might not be accurate.In order to process the metrical data and find a parametrical estimation method with the ability of excluding error interference,this study introduces a robust estimation into data pre-processing and establishes an iteration method with variable weights.The method proposed enables the problem be solved efficiently within the scope permitted.Furthermore,the study proves the effectiveness of the method proposed with specific project examples.