稳健估计在沉降监测数据处理中的应用

Application of Robust Estimation in Settlement Monitoring Data Processing

针对实际测得沉降观测数据中存在多种误差,严重影响观测结果的可靠性。由于最小二乘估计的方法没有抗差能力,所以引入稳健估计的方法对数据进行处理,使得观测值尽可能避免粗差的影响,得到最佳的估值。通过运用MATLAB软件建立回归方程,并对比分析了数据中含有粗差时2种方法的优越性。结果表明:监测数据中无粗差时,2种估计方法基本一致无明显差异,都可用于沉降监测数据处理;当数据中含有粗差时,稳健估计明显优于最小二乘估计,具有良好的抗差性,可以用于沉降监测数据处理。

The actual measured settlement observation data will inevitably have a variety of errors, and the existence of gross errors will greatly affect the reliability of the observation results. Since the method of least squares estimation has no resistance, the method of robust estimation was introduced to process the data so that the observations avoid the influence of gross errors as much as possible and obtain the best estimate. The regression equation was established by using MATLAB software, and the superiority of the two methods was compared and analyzed when the data contains gross error. The results show that there is no obvious difference between the two estimation methods when there is no gross error in the monitoring data, and they can be used for settlement. Monitoring data processing;when the data contains gross errors, the robust estimation is significantly better than the least squares estimation, has good resistance, and can be used for settlement monitoring data processing.