6R焊接机器人逆解算法与焊接轨迹误差分析

发布时间：2019-02-11 08:08

【摘要】：为了提高6R焊接机器人的位姿精度和焊接轨迹的准确度,提出了一种基于RBF神经网络的6R焊接机器人逆运动学求解方法。针对6R焊接机器人逆运动学方程组具有高维、非线性、求解复杂的特点,基于RBF神经网络建立运动学逆解预测模型,采用尺度空间理论对焊接机器人的位姿参数样本所在的工作空间进行分区,采用均匀设计法和模糊聚类理论对分区后的训练样本进行优选,并根据Z-Y-Z坐标转换原理进行转换和归一化处理,将逆运动学求解问题转换为基于RBF的6输入6输出预测系统。运用该系统对6R焊接机器人进行了复杂焊接轨迹仿真和点焊实验,并与基于组合优化迭代法和BP神经网络的逆运动学求解效果与焊接精度进行了比较,结果表明,基于RBF的6R焊接机器人运动学逆解预测模型具有求解简单、精度高、便于轨迹规划的特点,证明了该方法的可行性和有效性。
[Abstract]:In order to improve the accuracy of position and pose of 6R welding robot and the accuracy of welding trajectory, a method of inverse kinematics of 6R welding robot based on RBF neural network is proposed. The inverse kinematics equations of 6R welding robot have the characteristics of high dimension, nonlinearity and complex solution. The prediction model of inverse kinematics solution is established based on RBF neural network. The workspace of position and pose parameter samples of welding robot is partitioned by scale space theory, and the training samples after partition are selected by uniform design method and fuzzy clustering theory. According to the principle of Z-Y-Z coordinate transformation, the inverse kinematics problem is transformed into a 6-input-6-output prediction system based on RBF. The complex welding trajectory simulation and spot welding experiment of 6R welding robot are carried out by using this system, and the results are compared with the inverse kinematics solution and welding precision based on combinatorial optimization iterative method and BP neural network. The inverse kinematics prediction model of 6R welding robot based on RBF has the advantages of simple solution, high precision and convenient trajectory planning. The feasibility and effectiveness of this method are proved.
【作者单位】：四川大学制造科学与工程学院;桂林航天工业学院机械工程学院;南安普顿大学环境与工程学院;
【基金】：“十二五”国家科技支撑计划项目(2015BAF27B01) 四川省科技计划项目(2015GZ0036、2016GZ0195) 广西高校中青年教师基础能力提升项目(KY2016YB535) 广西高校机器人与焊接重点实验室主任基金项目(JQR2015ZR04)
【分类号】：TP183;TP242

【相似文献】