高文斌, 王洪光, 姜勇, 潘新安. 基于距离误差的机器人运动学参数标定方法[J]. 机器人, 2013, 35(5): 600-606. DOI: 10.3724/SP.J.1218.2013.00600
引用本文: 高文斌, 王洪光, 姜勇, 潘新安. 基于距离误差的机器人运动学参数标定方法[J]. 机器人, 2013, 35(5): 600-606. DOI: 10.3724/SP.J.1218.2013.00600
GAO Wenbin, WANG Hongguang, JIANG Yong, PAN Xin'an. Kinematic Calibration Method of Robots Based on Distance Error[J]. ROBOT, 2013, 35(5): 600-606. DOI: 10.3724/SP.J.1218.2013.00600
Citation: GAO Wenbin, WANG Hongguang, JIANG Yong, PAN Xin'an. Kinematic Calibration Method of Robots Based on Distance Error[J]. ROBOT, 2013, 35(5): 600-606. DOI: 10.3724/SP.J.1218.2013.00600

基于距离误差的机器人运动学参数标定方法

Kinematic Calibration Method of Robots Based on Distance Error

  • 摘要: 根据指数积公式中关节旋量坐标的理论值和实际值之间的伴随变换关系,将机器人运动学方程改写成包含有关节约束条件的等价形式.对运动学方程取微分得到末端连杆定位误差与关节旋量误差及零位位置误差间的线性化模型.利用距离精度的定义给出了指数积形式的机器人距离误差模型,避免了标定过程中坐标测量结果从测量设备坐标系向机器人基础坐标系变换带来的误差.基于最小二乘法给出了一种机器人运动学参数标定模型.采用激光跟踪仪作为测量手段完成了一种5自由度机械臂运动学参数标定实验.测试结果表明,经过参数标定在随机产生的测试点上机器人距离误差的平均值降低了近3.5倍.

     

    Abstract: According to the adjoint transformation relationship between the theoretical and actual values of joint twists in the product of exponential (POE) formula, the robot's kinematics equation is transformed into an equivalent form including the joint constraints. The linearized equation describing the relationship between the positioning errors of the end-link and the errors in the joint twists and in the zero position is obtained by differentiating the kinematic equation. With the definition of distance accuracy, the robot's distance error model is set up in POE form, which can avoid the coordinate transformation error from measurement equipment coordinate system to robot's base coordinate system. A kinematic calibration model of robots is presented based on the least-squares method. Finally, a kinematic parameters calibration experiment of a 5-DOF (degree of freedom) manipulator is completed with the help of a laser tracker. The testing result shows that the mean value of the distance errors of the robot is reduced about 3.5 times on the testing points generated randomly after parameter calibration.

     

/

返回文章
返回