关节式坐标测量机参数识别算法研究

摘要

In order to select the best one from the three common calibration algorithms, namely, Nonlinear Least Squares Method, Genetic Algorithm and Simulated Annealing Algorithm, which are used to identify the parameters of AACMM (articulated arm coordinate measuring machine), a unified loss function was defined and the experimental data were divided into the calibration group and test group. A comparative analysis was made in terms of an algorithm's speed, effectiveness and stability. The results show that LM method as one of Nonlinear Least Squares Methods is very suitable for parameter identification of AACMM as it is the least time-consuming and produces the most stable outputs among the three algorithms tested.%为了对比关节式坐标测量机结构参数常见的三种标定算法:非线性最小二乘法、遗传算法和模拟退火算法,建立了统一的损失函数,将实验数据分为标定数据和测试数据,分别用LM法、遗传算法、模拟退火算法标定关节式坐标测量机,从算法的速度、实测效果和稳定性三方面进行了对比分析.结果表明:LM法耗时最短,所得结果也最为稳定,是适合进行关节式坐标测量机标定的优良算法.