双步抗差估计及其应用

Two-step Robust Estimation and Application

抗差估计是一种可以有效抵制粗差影响的平差方法。但权因子函数的临界值定值问题较为困难,若选取不当,会造成法方程系数阵秩亏或粗差观测值的影响未能完全消除,从而导致平差无法进行或抗差不彻底。针对上述问题提出了一种改进型抗差估计法—双步抗差估计法,并通过对一边角网和一实测GPS向量网进行的试算,验证了双步抗差估计的优越性。

Robust estimation is an effective adjusting method of resisting gross error. But it is difficult to determine the critical value of weighting factor function. It could make the rank defect of normal equation coefficient matrix and incomplete removal of gross error influence for inappropriate choice of the critical value of weighting factor function, which result unable adjusting or halfway robust estimation. For these problems, in this paper, two-step robust estimation which is a improved robust estimation is proposed and the advantage is verified through the trial of a triangulation network and actual GPS vector network.