基于MDH模型的工业机器人运动学标定技术的研究

发布时间：2018-07-03 21:15

本文选题：工业机器人 + 标定　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：随着工业机器人的应用范围越来越广泛,人们对工业机器人的精度也提出了更高的要求。工业机器人的重复定位精度高,但其绝对定位精度低的特点限制了其在一些领域中的应用。标定技术是目前提高工业机器人绝对定位精度的最有效的方法之一。本文通过对工业机器人运动学标定技术进行研究,旨在提高工业机器人的绝对定位精度。本文以IRB4600型工业机器人为研究对象,首先根据D-H建模方法,建立了工业机器人的运动学模型,进行了正运动学和逆运动学的推导,采用Robotics工具箱对所建立的运动学模型进行了仿真验证。对影响机器人末端定位精度的因素进行了分析,针对D-H模型在平行轴处的奇异性和参数突变性的问题,本文采用MDH建模方法在平行轴处建模,然后基于微分变换的思想推导出了机器人的定位误差模型,采用最小二乘法进行参数辨识,采用预补偿的方式对机器人进行误差补偿。误差模型中存在的冗余性参数会对参数辨识的准确性产生很大的影响。针对这一情况,首先在理论上分析了误差模型中的冗余参数对参数辨识精度的影响,然后采用Khalil提出的误差分析方法,通过对相邻关节几何关系的推导,分析了不同结构的串联机器人位置误差模型中的冗余性参数。最后以IRB4600型工业机器人为对象对冗余性参数的分析进行了仿真验证。在机器人的整个工作空间内采集标定所需的数据,对机器人进行了标定实验。实验结果表明,经过标定机器人的绝对定位精度提高了89.75%,机器人的末端绝对定位误差可保证在0.5mm以内。通过对比采集点处去除冗余参数和未去除冗余参数的误差补偿结果,验证了冗余参数分析的正确性。另外对机器人进行了较小的局部工作空间的标定实验,通过在局部工作空间内和局部工作空间外采集测试点,来对比两次不同范围标定的误差补偿效果,结果显示通过缩小标定空间,可以进一步提高机器人在标定空间内的绝对定位精度。
[Abstract]:With the wide application of industrial robots, the precision of industrial robots is required more and more. The repeated positioning accuracy of industrial robot is high, but its low absolute positioning accuracy limits its application in some fields. Calibration technology is one of the most effective methods to improve the absolute positioning accuracy of industrial robots. In this paper, the kinematics calibration technology of industrial robot is studied to improve the absolute positioning accuracy of industrial robot. In this paper, IRB4600 industrial robot is taken as the research object. Firstly, according to D-H modeling method, the kinematics model of industrial robot is established, and the forward kinematics and inverse kinematics are deduced. The kinematics model is simulated by Robotics toolbox. The factors that affect the precision of robot terminal positioning are analyzed. In view of the singularity of D-H model at the parallel axis and the problem of parameter mutation, the MDH modeling method is used to model the parallel axis. Then, based on the idea of differential transformation, the positioning error model of the robot is derived. The least square method is used to identify the parameters, and the pre-compensation method is used to compensate the robot error. The accuracy of parameter identification is greatly affected by redundant parameters in the error model. In order to solve this problem, the influence of redundant parameters in error model on the accuracy of parameter identification is theoretically analyzed, and then the geometric relationship of adjacent joints is deduced by using the error analysis method proposed by Khalil. The redundant parameters in the position error model of series robot with different structures are analyzed. Finally, the redundant parameters of IRB 4600 industrial robot are simulated and verified. The calibration data are collected in the whole workspace of the robot and the calibration experiment is carried out. The experimental results show that the absolute positioning accuracy of the calibrated robot is improved by 89.75, and the absolute positioning error of the robot can be guaranteed within 0.5mm. The correctness of redundant parameter analysis is verified by comparing the error compensation results of removing redundant parameters and not removing redundant parameters at acquisition points. In addition, the small local workspace calibration experiment is carried out, and the error compensation effect of two different calibration ranges is compared by collecting the test points in the local workspace and outside the local workspace. The results show that the absolute positioning accuracy of the robot in the calibration space can be further improved by reducing the calibration space.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242.2

【参考文献】