摘要
针对一种3P3R型串联机器人,建立了参考零位模型与DH(Denavit-Hartenberg)模型的混合运动学模型,将直线运动部分与旋转运动部分分开建模,能够更好地描述机器人不同机械结构的几何关系,在此基础上提出了结合几何辨识和参数辨识的两步标定方法.然后,结合机器人的机械结构特点,分析了机器人在操作大型零件过程中的结构变形,并提出了考虑结构变形的运动学补偿模型.最后,使用激光跟踪仪完成了机器人标定实验,通过对比空载和加载情况下的定位误差,验证了运动学标定和补偿的效果.结果表明,混合运动模型采用两步参数辨识能够在空载情况下取得较高的标定精度,而运动学补偿模型则能够在加载情况下对运动学进行较好的变形误差补偿.
A hybrid kinematic model consisting of zero-reference position model and DH （Denavit-Hartenberg） model is proposed for a 3P3R structured robot. The linear motion part and the rotational motion part are separated in this model for better description of geometric relationship among different mechanical structures. And a two-step calibration method with geometric and parametric identification is proposed for this model. Then the structural deformation of the robot during manipulating the large scale part is analyzed with the structural features of the robot, and a kinematic compensation model considering structural deformation is proposed. Finally, a calibration experiment is conducted with a laser tracker, the effects of the kinematic calibration and compensation are evaluated by comparing the positioning errors of no-load operation and loaded operation. The results show that the hybrid kinematic model with two-step parameter identification method works well for the no-load operation and that the compensation model can provide a good compensation for kinematic errors due to deformation under heavy load.
出处
《机器人》
EI
CSCD
北大核心
2015年第3期376-384,共9页
Robot
基金
国家自然科学基金资助项目(51105218)
关键词
机器人标定
运动学补偿
激光跟踪仪
结构变形
robot calibration
kinematic compensation
laser tracker
structural deformation
