Research On 3D Laser Guided Robotic Welding Trajectory Planning For Bicycle Tripod

At present,industrial robots have been widely used in the field of welding,especially for the welding of standard workpieces.However,for the welding of non-standard workpieces,due to the difference between the individual workpieces,the teaching-based robot welding scheme cannot sense the difference in the weld position between the actual workpiece and the teaching workpiece,resulting in deviation of the welding precision.Vision-guided robotic welding can sense the actual workpiece weld location and improve welding accuracy,and has become an urgent need of manufacturing companies.There are many researches on vision-guided robotic welding,mainly in flat simple welds,but less on scenes with complex space welds and high precision requirements.Therefore,this paper proposes an automatic welding scheme based on visual scanning and weld reconstruction for complex non-standard workpiece welding scenarios.In this paper,several key technologies involved in the robotic welding system of complex space curves: sensor pose calibration,joint position planning and trajectory planning.First,the transformation relationship between the coordinate system of the robot workstation is determined in the vision-based robotic welding system.And since the vision sensor is fixed at the end of the robot,it is necessary to calibrate the transformation matrix of the sensor relative to the end of the robot to determine the positional relationship of the weld relative to the base coordinate system of the robot.Through the research of the domestic and international sensor calibration methods and various applicable scenarios,combined with the line structure light sensor and current experimental conditions,the three-point calibration scheme of structured light sensor is designed.At last,the inverse matrix solution in the formula is eliminated to make the computer program more convenient to implement the algorithm.In addition,this paper designs a sensor continuous scanning and robot position interpolation scheme,determines the robot end position when the sensor scans to each weld point,and transforms the weld position from the sensor coordinate system to the robot base coordinate system.After determining the position of the weld,the robot welding pose was planned.Firstly,the weld coordinate system and the welding gun coordinate system are established based on the welding process parameters,then the transformation matrix of the welding gun coordinate system relative to the robot base coordinate system is obtained.Since the welding gun rotation angle can be arbitrarily set,the welding space dimension is smaller than the robot motion dimension,which leads to the redundancy of the six-axis robot,and the attitude optimization scheme is proposed.The attitude optimization scheme optimizes the motion performance of the robot during welding by setting the angle of the welding gun.Firstly,the Kinematics performance evaluation model is established,and then based on the the model,combined with the genetic algorithm to find the rotation angle corresponding to the best Kinematics performance.The position and attitude planning of the welding point is completed,thereby obtaining the sequence of postures required for the robot welding.After obtaining the point pose sequence,A smooth welding track need to plan.In this paper,the NURBS-based robotic welding trajectory planning scheme is designed for the problem that the current trajectory planning method does not pass the path point and the welding precision is reduced,and the welding trajectory is not smooth.Firstly,the formula of NURBS curve is deduced,and how to solve the control point through the path point is analyzed.Then,the NURBS curve is used to plan the trajectory of the robot joint space.This paper also uses the C++ programming language to implement the NURBS trajectory planning algorithm.Then,the architecture and scheme flow of the robotic welding system are designed as a whole.The hardware selection and software design are respectively carried out for each subsystem.The subsystems are the upper computer control subsystem,the robot and welder execution subsystem,and the visual sensing subsystem.And Socket communication is used to connect subsystems.This paper designs an acknowledgment feedback mechanism based on Socket communication to ensure the accuracy of information transmission.Finally,based on the existing welding station and the automatic welding platform designed by this subject,the algorithm introduced in the paper was verified,and the independent welding physics experiment of the bicycle tripod weld was completed.The simulation and physical experiments respectively verified the communication between the robot and the host computer,sensor calibration,pose planning,trajectory planning,etc.Finally,all the modules were integrated to complete the welding of the tripod weld.The welded joint surface after welding is smooth,and the complete coverage of the weld meets the accuracy requirements,which proves the effectiveness and practicability of the 3D vision-guided robotic welding platform designed in this paper.