常用稳健估计方法应用于水准网的相对增益比较

Compare the Relative Gain of Commonly Used Robust Estimation Methods for Leveling Networks Calculation

稳健估计方法常用来消除或减弱观测值中的粗差对水准网计算的影响,然而,不同稳健估计方法的稳健性不尽相同。以3个具有不同观测值数量和未知数数量的水准网为例,通过仿真实验比较了13种常用稳健估计方法的稳健性。结果表明,无论是对于具有独立等权观测值的水准网还是对于具有独立不等权观测值的水准网,L1法、Danish法、German-McClure法和IGGIII方案都是相对更为有效的稳健估计方法。它们比其他常用稳健估计方法能更有效地消除或减弱粗差的影响。

Robust estimation methods are often used to eliminate or weaken the influences of gross errors on level-ing networks calculation. However, different robust estimation methods may have different capabilities in elimina-ting or weakening gross errors. Taking 3 leveling networks with different number of observations and unknowns as examples, the simulation experiments have been employed to compare the robustness of 13 frequently used robust estimation methods in this paper. The results indicate that no matter the independent elevation difference observa-tions have equal weights or unequal weights, the L1, Danish and German-McClure methods and IGGIII scheme are relatively more efficient among the 13 commonly used robust estimation methods. These four methods can efficiently eliminate or weaken the influences of gross errors on leveling networks calculation.