Research On Robot Constant Force Polishing Technology Of Large-Scale Curved Workpiece Based On 3D Vision

In recent years,although robot technology has been vigorously developed,the manual way of grinding and polishing the workpiece has been used,especially for large curved surface workpieces.Because of the complexity of programming and the difficulty of obtaining model data,it is difficult to realize the automatic grinding and polishing of the workpiece,which seriously hinders the intelligent development of China’s manufacturing industry.The development of 3D vision technology provides a new way for the automatic grinding and polishing of robots.Aiming at the problems existing in automatic grinding and polishing,this paper studies the robot constant force grinding and polishing technology of large curved surface workpiece based on 3D vision technology.Taking the large curved surface head as the research object,the constant force control of the grinding process,the splicing of the workpiece point cloud data,the path planning of the workpiece grinding and the automatic programming of the robot are realized,and then the automatic grinding and polishing of the workpiece are realized.The main research contents of this paper are as follows:A kind of end effector used for grinding and polishing is designed.The zero point calibration and gravity compensation algorithm of the end effector and the force control algorithm based on PID are proposed.Firstly,the force analysis of the end-effector was carried out,and the mathematical relationship between the surface pressure of the workpiece and the measured force was obtained.Then,the gravity compensation in any position and orientation of the robot was realized according to the proposed zero calibration and gravity compensation algorithm.Aiming at the force control problem of robot grinding and polishing,a force control algorithm based on PID is proposed,and the force signal is filtered by the first order inertial filter.The experimental results show that the actual force error is within plus or minus 4N,and the filtered force error is within plus or minus 0.5N.Based on Preston’s empirical equation,the material removal model of circular tool is established and the algorithm for determining the optimal row spacing is proposed.Firstly,the material removal models of the circular tool were established under the straight path and the curve path respectively.Then,under the same conditions,the material removal model of the circular tool was obtained by subtracting the material removal models of two circular tools with different radii.In addition,based on the contour map of the material removal model,an algorithm for determining the optimal row spacing was proposed.A point cloud data splicing algorithm based on robot pose information is proposed and the point cloud data is preprocessed.Firstly,the point cloud data of each part of the workpiece is collected by a 3D camera,and the collected point cloud is spliced by using the pose information of the robot to obtain the complete point cloud of the workpiece.Then,the point cloud data is preprocessed by voxel filtering and pass-through filtering,which reduces the point cloud density,improves the point cloud quality,and reduces the processing time of the subsequent algorithm.A path planning algorithm for grinding and polishing based on NURBS curve fitting and polynomial curve fitting is proposed.Firstly,the path point cloud is extracted according to the defined plane parameters,and the NURBS curve is fitted to the path point cloud.After the NURBS curve is discretized,the coordinates of the grinding point are obtained,and then the tangent vector and normal vector of the grinding point are estimated.The path planning of workpiece grinding and polishing also needs to determine the adjacent path position.Firstly,the normal point cloud of the path was extracted according to the proposed algorithm,and the normal point cloud was fitted by polynomial fitting.Then the position of the adjacent path was determined according to the determination algorithm of the adjacent path position.The robot constant force grinding and polishing system is designed and built,and the experimental analysis of the proposed algorithm is carried out.The hardware of the grinding and polishing system is composed of industrial robot,end-effector,industrial personal computer,3D camera,etc.The software mainly includes point cloud processing interface,host computer interface,Codesys program interface,etc.