| As important equipment for the intelligent transformation of the manufacturing industry,industrial robotic arms have become a research hotspot in the field of robotic arms on how to integrate information from external sensors.The integration of machine vision technology and robotic arms can effectively improve the intelligence level of robotic arms in industrial production,but the development of integrated control systems has problems such as a long time,large workload,and high research and development costs.To effectively promote the intelligence of robotic arms,shorten the integrated research and development cycle,and reduce development costs,this thesis introduces the concept of digital twins into the integrated research and development of robotic arms and machine vision.The UR5 robotic arm grasping obstacle avoidance is the object of research,which is carried out from machine vision technology,integrated control system virtual debugging,and integrated control system transition to real space.The specific research of the paper is as follows:(1)Research on key technologies for obstacle avoidance in visual-based robotic arm grasping.Key technologies include building an integrated control framework for robotic arms based on ROS,target recognition and positioning algorithm YOLOv5,obstacle spatial position acquisition algorithm point cloud filtering processing,and hand-eye calibration for determining the position of depth cameras and robotic arms in real space.(2)Research on virtual commissioning of UR5 robotic arm grasping and obstacle avoidance.Select the virtual space Gazebo,build an experimental environment for virtual debugging of UR5 robotic arms,and establish Move It! Communication connection with virtual UR5 robotic arm;Test the reliability of the object detection algorithm YOLOv5 and point cloud filtering processing in virtual space;Visual-based virtual debugging for robotic arm grasping and obstacle avoidance.(3)Research on obstacle avoidance of UR5 robotic arm grasping in real space.Establishing an experimental environment in real space for UR5 robotic arms to grasp and avoid obstacles;Conduct hand-eye calibration experiments to obtain the coordinate relationship between the depth camera and the UR5 robotic arm;Collect the obstacle position data in the working space of UR5 manipulator,debug the point cloud filtering processing and YOOv5 target detection,drive the virtual UR5 manipulator with real data,and carry out the experiment of grasping,placing and motion planning of grasping and obstacle avoidance;Move It! Establish contact with the real UR5 robotic arm and conduct research on the transition of a vision-based integrated control system from virtual space to real space. |
