摘要
为了有效地提高关节式坐标测量机的测量精度,变臂关节式坐标测量机通过锁定坐标测量机的前2个关节,减小前端关节转角误差的引入降低系统传递误差。经D-H方法建模,变臂关节式坐标测量机和关节式坐标测量机具有相同的数学方程,但是两者具有不同的结构参数,主要差别体现在前2个关节转角。提出了一种基于逆运动学分析方法和非线性规划遗传算法的参数自标定方法。首先,采用Robotics Toolbox中的逆运动学分析方法反求到达同一标定点的多组可能位姿;其次,根据这些反求的位姿结合测量模型计算标定点的坐标值;最后,以单点重复误差为目标函数通过遗传算法进行全局搜索并利用非线性规划算法进行局部搜索获取最优标定结果。实验结果表明,所提出的参数自标定方法使得变臂关节式坐标测量机的单点测量精度提高1 000倍,长度测量精度提高8倍。
In order to effectively improve the measurement precision of partly-bonded articulated arm coordinate measuring machine （AACMM）,the system propagation error of the AACMM is reduced by bonding the front two articulated arms of the coordinate measuring machine and fixing the two front joint angles.The D-H modeling method indicates that the partly-bonded AACMM and the AACMM have the same mathematical equation,but they are different in structure parameters and their main difference focuses on the front two joint angles.This paper presents a parameter self-calibration method based on the inverse kinematics analysis method and genetic nonlinear programming algorithm.Firstly,the inverse kinematics analysis method in Robotics Toolbox is used to inversely acquire several groups of the possible poses for the same calibration point ; Secondly,the coordinates of the calibration point are calculated according to the obtained poses and the measurement model; Finally,with minimizing the single point repeatability error as the objective function,the optimal calibration results are obtained using local search with the nonlinear programming algorithm and global search with the genetic algorithm.The experiment results show that the presented parameter self-calibration method improves the single point measurement precision of the partly-bonded AACMM by 1 000 times,and improves the length measurement precision of the partly-bonded AACMM by 8 times.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2014年第3期572-579,共8页
Chinese Journal of Scientific Instrument
基金
中央高校基本科研业务费资金(12MS122
13XS35)资助项目
关键词
变臂关节式坐标测量机
参数标定
D-H方法
非线性规划遗传算法
逆运动学分析
partly-bonded articulated arm coordinate measuring machine
parameter calibration
D-H method
genetic nonlinear programming algorithm
inverse kinematics analysis
