The nonlinear links such as dead zone and gap in the joints of the manipulator of the welding robot will complicate the dynamic behavior of the system.There is a motion coupling relationship between each joint and the actuator,which leads to nonlinear jitter of the joint in the local area,resulting in control errors.For this reason,a nonlinear dual-arm jitter control method for complex metal joint welding robot is proposed.The motion coordinate system of welding robot is constructed,the transformation matrix of terminal robot is obtained,and the complex nonlinear motion coupling analysis of both arms joint is realized.Consider-ing the existence of nonlinear links such as dead zone and clearance in joint operation,a linear function is in-troduced to output the actual rotation angle of the joint.The rotation angle is taken as the input of the radial basis function neural network,and Lagrange's law is used to transform the conversion matrix into the angular displacement,angular velocity and angular acceleration of the robot arm,which are introduced into the non-linear integral sliding mode governing equation to compensate the friction torque,centrifugal force matrix and gravity of the robot arm,so as to reduce the nonlinear jitter,and finally obtain the adaptive control result.Thus the control error is reduced.The experiment shows that the proposed method can deal with the inherent uncertainty,nonlinearity and external interference in the welding process,improve the welding accuracy and stability,and help to improve the welding qualityand consistency.
关键词
焊接机器人/金属焊接/运动学模型/径向基函数神经网络算法
Key words
welding robot/metal welding/kinematic model/radial basis function neural network al-gorithm
维普
万方数据