串联机器人加工系统关节刚度的高精度辨识方法

提出一种串联机器人加工系统关节刚度的高精度辨识方法.首先,基于雅克比矩阵的Frobenius范数,得到串联机器人加工系统的灵巧性,并确定工作空间内可达姿态的灵巧性数值分布情况.然后,分别选取4组不同等级灵巧性的姿态进行关节刚度辨识实验,并计算出相应姿态下的关节刚度.最后,对辨识出的4组关节刚度的末端变形计算值与测量值进行相对误差分析.结果表明:随着灵巧性的增大,机器人关节刚度辨识的准确性越高;相较于灵巧性较小的姿态,灵巧性较大的姿态的末端变形计算值与测量值的相对误差可降低20%~50%.

High Accuracy Identification Method of Joint Stiffness in Serial Robot Machining System

A high-precision joint stiffness identification method for a serial robot machining system was proposed. First, based on the Frobenius model number of the Jakobi matrix, the kinematic conditioning index of system was obtained and the numerical distribution of kinematic conditioning index at the robot accessible postures in the working space was determined. Then, four groups of postures with different levels of kinematic conditioning index were selected to carry out the joint stiffness identification experiment, and the joint stiffness was calculated under the corresponding posture. Finally, the relative error between the end deformation calculation value and measurement value of the identified four groups of joint stiffness was analyzed. The results show that the accuracy of robot joint stiffness recognition increases with the increase of kinematic conditioning index, and the relative error is reduced by 20%-50% compared to that at the posture with less kinematic conditioning index.