基于RANSAC算法的工业机器人运动学参数标定  

Kinematic parameter calibration of industrial robot based on RANSAC algorithm

作　　者：邹方星 范百兴[1] 陈哲[1] ZOU Fangxing;FAN Baixing;CHEN Zhe(School of Geospatial Information,Strategic Support Force Information Engineering University,Zhengzhou 450001,China)

机构地区：[1]信息工程大学地理空间信息学院,郑州450001

出　　处：《测绘工程》2024年第4期9-15,共7页Engineering of Surveying and Mapping

摘　　要：为提升工业机器人末端绝对定位精度,文中提出一种基于RANSAC算法的运动学参数标定算法。首先采用D-H模型建立末端位置运动学误差模型;通过RANSAC算法对标定点集进行筛选,剔除误差较大的粗差点;最后采用抗差岭估计辨识运动学参数,减少测量扰动、粗差和法方程病态性等因素的影响。试验结果表明,机器人绝对定位精度的RMS误差由0.999 mm降低至0.229 mm,误差下降77.08%。相比于传统的标定算法,文中算法辨识运动学参数准确性更好,标定精度更高。In order to improve the absolute positioning accuracy of industrial robots,it is proposed a kinematic parameter calibration algorithm based on RANSAC algorithm.Firstly,the kinematic error model of the end position is established by using D-H model.Then,it is screened by RANSAC algorithm of the set of standard points,and the rough handicap with large error is eliminated.Finally,the kinematic parameters are identified by the tolerance ridge estimation,and the influence of measurement disturbance,gross error and ill-condition of the normal equation is suppressed.The experimental results show that the RMS error of the absolute positioning accuracy of the robot is reduced by 77.08%from 0.999 mm to 0.229 mm.Compared with the traditional calibration algorithm,the proposed algorithm is more accurate and accurate in identifying kinematics parameters.

关 键 词：运动学参数标定 随机采样一致算法 抗差岭估计 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]