Research On Real-time Welding Seam Tracking Control System Of Welding Robot

Based on the theory and method of trajectory optimization of particle swarm optimization and combined with visual recognition technology,this paper designs and proposes a set of weld tracking control system of welding robot,in order to solve the problems of low precision,poor weld forming quality and high labor intensity in the manual teaching welding process.It mainly carries out three aspects:trajectory planning of welding robot,image recognition processing and experimental implementation of tracking system:Taking the Anchuan ma1440 welding robot as the research object,the D-H model is established,and the forward and inverse kinematics solutions are analyzed.Using the robotics toolbox module,the mathematical model of the robot is constructed in MATLAB,and the forward and inverse motions of the robot are simulated respectively to verify the correctness of the forward and inverse kinematics algorithms.Based on the kinematic model,the cubic and quintic polynomial interpolation method in joint space and the line and arc interpolation method in Cartesian space are studied.Through the simulation verification on MATLAB platform,the simulation results are obtained.By comparing and analyzing the time optimal piecewise polynomial trajectory planning based on particle swarm optimization with the trajectory planning based on 3-5-3 polynomial interpolation method,it can be concluded that particle swarm 3-5-3 piecewise polynomial interpolation trajectory planning can ensure the stability of robot man in the process of weld tracking and shorten the running time by 5.7 seconds.Finally,the robot is programmed offline based on motosimeg VRC software,and the program is imported into the robot to perform the weld tracking task according to the planned trajectory.Gray processing and filtering algorithm are used to preprocess the weld image to eliminate the image interference.An edge operator detection method is proposed to extract the edge information of the weld area.The morphological calculation method is used to fill the extracted weld contour and smooth the edge of the weld contour.Aiming at the problem of extracting the weld centerline,the template matching algorithm is applied to extract the accurate position coordinates of the weld centerline.The weld tracking control platform of welding robot based on image processing is designed and realized,the laser vision sensor is calibrated,the coordinate conversion between robot and vision sensor is established,and the communication is completed.In order to realize the real-time control of weld tracking,the host computer software is designed to realize the management of weld image processing and robot trajectory coordinate control.The experimental research shows that the weld tracking control system of welding robot is more accurate than manual teaching,the robot has continuous trajectory and short time-consuming.It has reference significance for the control of key processes of welding production in small and medium-sized enterprises.