Research On Visual Guidance And Control Method Of Welding Robot

Traditional welding robot machining methods rely heavily on teaching or off-line programming which requires high positioning accuracy of fixture,resulting in low flexibility and poor adaptability in the process of production.In order to increase the flexibility of the welding robot processing mode,remove the complex positioning device and achieve the goal of robot autonomous tracking of welding seam,this dissertation studies the welding seam visual guidance system and its motion control method.Here are several aspects of work:The overall scheme of the visual guidance system is determined,and the eye-inhand configuration is adopted.The mathematical expressions of camera imaging models along with the internal and external parameters are studied.The calibration of camera parameters is performed by using Zhang Zhengyou's calibration method,and the average error is 0.054 pixel.The calibration method of structured light plane is studied.The coordinate values of light points on two groups of parallel light stripes are obtained by setting transition coordinate system.The parameters of light plane are calibrated by SVD method.The calibration of hand eye matrix is completed,and the effect of the number of calibrated images on the calibration accuracy is explored.The conclusion is that the calibration accuracy can be ensured when the number of calibration images is about 15.The D-H parameters of the articulated arm robot are modeled,and the forward and inverse kinematics problems of the robot are derived.By using the differential meaning of Jacobian matrix,the inverse kinematics problem is solved numerically.In order to get the desired orientation,the unit quaternion is used to interpolate the orientation changes,so that the pose of the end-effector of robot can change smoothly.An image processing algorithm for obtaining the center point of welding seam is completed,and the point obtained by visual sensor is to be interpolated.The trajectory of the robot is planned by using the hybrid trajectory between the obtained joint angles to ensure the smooth motion of the six joints in the motion of the articulated arm robot.A user interface which integrates image processing to obtain welding seam central point,to calculate joint angle and real-time control of robot is programmed.A complete system is obtained.The tracking error of straight line is about 0.4 mm.