基于稳健估计不同权函数抗差效果分析

Analysis of Different Weight Functions Based on Robust Estimation

人为使用测量设备获取数据时,数据中可能存在误差不能达到绝对的准确及有效,使用不同稳健估计方法达到稳健估计效果不同。以重庆草堂隧道为研究区,以中部斜井地表观测点沉降为研究对象,当数据中含有粗差时,稳健估计的方法相比于最小二乘法更能达到理想效果,但是不同的稳健估计方法达到的效果也会不同。通过对实测数据进行稳健估计研究,得出Huber权函数相比于其他权函数能更好地剔除粗差,且误差更小。

When the measurement equipment is used to obtain the data manually,there may be errors in the data that cannot achieve absolute accuracy and effectiveness.The Caotang Tunnel in Chongqing was taken as the research area,and the settlement of the surface observation point of the inclined shaft in the middle was taken as the research object.When the data contains coarse error,the method of robust estimation can achieve the ideal effect better than the least square method,but different methods of robust estimation can also achieve different effects.Based on the robust estimation of the measured data,it is concluded that the Huber weight function can eliminate the gross error better than other weight functions.