Research On Control System Of Grinding Robot Based On ROS

At present,manual grinding is still the main way to process small workpieces.In order to improve the automation and product quality,more and more researchers are attracted by robot grinding technology.This research proposes a control system based on ROS for grinding manipulator.In order to improve the MoveIt trajectory planning ability,the traditional RRT algorithm is researched and improved.In order to meet the real-time requirements in the control of the grinding manipulator,a grinding system consisting of a grinding manipulator,a computer and a belt machine is designed.After analyzing the requirements,the control system is configured as: the EtherCAT industrial bus of the control system is used to realize the expansion of communication bandwidth;the Xenomai+Linux dual-core real-time system is set up on the computer to realize the real-time control of the grinding manipulator;the upper layer of the control system is built with the ROS control system,and the ROS controller based on the ros_control framework is configured for the grinding manipulator.Next,the URDF file of the grinding manipulator is written to establish a ROS simulation model for the grinding manipulator.Then a trajectory planning package is configured by MoveIt Setup Assistant for the grinding manipulator,and the move_group node is used as the interface of the grinding robot controller and various computing plugins.The KDL and FCL are discussed and tested in RViz;Next,Constructing the Gazebo dynamic physical simulation platform to verify the planning results,debuging PD parameters in the ROS abstract controller.The parameters lay the foundation for the robotic body control experiment.Two planning strategies of RRT and EST were analyzed and the control effects are compared.In views of the limitations of MoveIt in the robot's end-effector trajectory planning capability,the RRT planning process is improved: based on the NURBS curve of the workpiece,a parallel secant method is used to extract the motion trajectory and the pose of grinding.The compute_cartesian_path function is wrote,improves the trajectory planning ability of the RRT algorithm for grinding manipulator by adding the trajectory and pose information.At last,the control system and the planning algorithm are verified on the physical grinding manipulator.Firstly,a single joint experiment is performed on the manipulator to verify the effectiveness of the real time control system.Then,the planning algorithm verification test is performed.Finally chamfering grinding experiments were carried out.A series of experimental results demonstrate the effectiveness of the grinding control system and its motion planning capabilities.