6R型工业机器人精度分析与结构参数标定的研究

发布时间：2018-06-04 18:34

本文选题：6R型机器人 + 精度分析　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：随着工业化的发展,工业机器人已逐步成为核心装备,且越来越多的行业对其精度有了较高的要求,因此开展工业机器人的定位精度的研究尤其重要。本文以6R型工业机器人为对象,从运动学的角度,对误差源中确定性的结构参数误差和随机性的关节间隙误差进行定位误差分析。根据误差特性对结构参数误差进行运动学标定,通过闭环约束标定方式和基于约束模型的辨识方法,标定得到准确的参数值,提高6R机器人绝对定位精度。利用几何法和蒙特卡洛法,求解出机器人工作空间;利用正逆运动学,分析机器人各结构参数对末端位置精度的影响,得出对机器人精度的影响规律。使用矢量法对关节间隙建模,基于概率论,分析机器人的二、三关节间隙对末端位置误差的影响,得到的误差分布能有效的指导机器人的设计、使用和维护,以保证其重复定位精度。基于微分运动学,建立考虑几何参数的6R型串联机器人的误差传递模型,并通过Matlab仿真验证参数误差模型的准确性。针对机器人闭环约束自标定方法,研究分析基于点、平面、球面约束的标定方法,提出一种新型的球心距离标定方法。分别对牛顿法、最小二乘法、下山单纯形法进行比较分析,找到最适合误差参数标定模型的辨识方法。针对参数的可辨识性,剔除无法辨识的参数,分析确定了17项待辨识参数。使用基于DETMAX的增减算法,对标定实验中的测量位姿进行筛选,使位姿信息参与误差辨识,提高参数辨识的准确度。对单球面约束标定实验进行仿真,使用Nelder-Mead算法进行参数辨识,补偿后位置点更趋于拟合球面,表明球面模型标定具有可行性。为了验证所提出的标定方法的有效性,针对不使用外部测量仪器的机器人自标定,采用平面约束和球心距离约束的方式,进行了两组标定实验。分别对采集到的测量点进行数据处理,通过下山单纯形法辨识得到结构参数偏差,在机器人控制器中对其进行补偿,得出实验结论。
[Abstract]:With the development of industrialization, industrial robot has gradually become the core equipment, and more industries have higher requirements for its precision, so it is particularly important to research the positioning accuracy of industrial robot. In this paper, a 6R industrial robot is used as an object. From the point of view of kinematics, the positional errors of deterministic structural parameters and random joint clearance errors in the error source are analyzed. According to the error characteristics, the kinematic calibration of the structural parameters error is carried out. Through the closed-loop constraint calibration method and the identification method based on the constraint model, the accurate parameter values are obtained and the absolute positioning accuracy of the 6R robot is improved. By using geometric method and Monte Carlo method, the workspace of robot is solved, and the influence of robot structural parameters on the precision of terminal position is analyzed by using forward and inverse kinematics, and the law of influence on robot precision is obtained. Using the vector method to model the joint gap, based on the probability theory, the influence of the second and third joint clearance on the end position error of the robot is analyzed. The error distribution can effectively guide the design, use and maintenance of the robot. In order to ensure its repeat positioning accuracy. Based on differential kinematics, the error transfer model of 6R series robot with geometric parameters is established, and the accuracy of the error model is verified by Matlab simulation. Aiming at the self-calibration method of robot with closed-loop constraint, the calibration method based on point, plane and sphere constraints is studied, and a new calibration method of spherical center distance is proposed. The Newton method, the least square method and the downhill simplex method are compared and analyzed respectively to find the most suitable identification method for the error parameter calibration model. In view of the identifiability of the parameters, 17 parameters are determined by eliminating the unidentifiable parameters. In order to improve the accuracy of parameter identification, an algorithm based on DETMAX is used to screen the measurement position and pose in the calibration experiment, so that the position and attitude information can participate in the error identification and improve the accuracy of parameter identification. The experiment of single sphere constraint calibration is simulated and the parameter identification is carried out by using Nelder-Mead algorithm. After compensation the position points tend to fit the spherical surface which indicates that the calibration of spherical model is feasible. In order to verify the effectiveness of the proposed calibration method, two sets of calibration experiments were carried out for robot self-calibration without external measuring instruments, using plane constraint and spherical center distance constraint. The measured points are processed separately and the deviation of structural parameters is obtained by using the downhill simplex method. The error of structure parameters is compensated in the robot controller and the experimental results are obtained.
【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】