Research On Robot Grinding System Based On Vision And Force Perception

With the progress of technology,industrial manufacturing is gradually developing from automation to intelligence.However,in the field of workpiece grinding,manual grinding is still the main method,which is not only inefficient,difficult to guarantee the quality,and not uniform in grinding standards,but also often produces harmful substances in the grinding process,which is harmful to human health.In view of the shortcomings of the current manual grinding method,this paper studies the robot grinding system,in order to improve the processing efficiency and product quality,improve the working environment of workers,and provide a new idea for automatic grinding.Firstly,this paper introduces the robot grinding control strategy based on force sensors,the current research status of force sensors and the vision-based target contour extraction method,and describes the main research contents of this paper.Secondly,the imaging model,perspective aberration,internal and external parameter matrix of pinhole camera and depth camera are derived,camera calibration and hand-eye calibration experiments are conducted,and the transformation matrix of camera and robot is obtained.Meanwhile,the 3D reconstruction of the target based on the vision system is studied and the point cloud is filtered.The edge extraction algorithm based on normal vector was performed for point cloud edge extraction,and the 3D spatial coordinates of the workpiece were obtained through the point cloud and handeye calibration matrix,and the spatial coordinates were transferred to the robot control system to ensure the accuracy of the robot to reach the grinding position,so as to realize the grinding operation.Then,in order to further improve the grinding accuracy,real-time acquisition of the contact force between the robot and the external environment based on the sixdimensional force sensor is carried out,and the parameter identification method is studied to eliminate the external interference of the contact force and is experimentally verified,and the results show the correctness of the solved parameters.At the same time,the force characteristics of robot grinding and robot force control method are studied and analyzed.An impedance algorithm based on position control is used to complete the conversion between contact force error and position error to achieve constant force control,and a force control simulation platform is built by simulation software to verify the feasibility of the algorithm,and the results show that the algorithm can effectively achieve constant force control in the robot grinding process.Finally,a robot grinding system based on force sensor and vision sensor is developed through the QT5 platform to realize the joint control of robot,depth camera and six-dimensional force sensor,and the reliability of the system for robot grinding operation is verified by conducting robot grinding experiments through the system.