常用稳健估计方法在测边网解算中的应用

Application of commonly used robust estimation methods in trilateration networks adjustment

针对最小二乘估计中稳健估计方法的效果尚未被分析证明的问题,该文以不同观测值数量、不同粗差数量和等权或不等权观测值的三个测边网为例,通过仿真实验的方法,对12种常用稳健估计方法的稳健性进行了比较,确定了对测边网解算相对更为有效的稳健估计方法。结果表明,L1法、Danish法、German-McClure法和IGGIII方案是12种方法中更为有效的稳健估计方法,能相对更有效地消除或减弱粗差对参数估计结果的影响。

Aiming at the problem that the efficiency of robust estimation methods in least square estimation has not been analyzed and proved yet,the paper took three different kinds of trilateration networks under the conditions of equal or unequal weighted observations with different numbers of observations and different numbers of gross errors as examples,compared the robustness between 12 commonly used robust estimation methods,and determined relatively more effective robust estimation methods for trilateration network calculation.Result showed that the algorithms of L1 method,Danish method,German-McClure method and IGGIII scheme among the 12 commonly used robust estimation methods could be more efficiently eliminate or weaken the influence of gross errors on the parameter estimation with their relative validity.