Research On AR-based Teaching Of 6 DOF Industrial Robot

Industrial robots are mechanical devices that perform repetitive work under program control.Due to their high flexibility and high precision,they have been widely used in various industrial production activities.At present,online programming and offline teaching are still the main methods of programming industrial robots.The method of online teaching programming process is clumsy and cumbersome;The method of offline teaching programming has high professional requirements,large modeling workload and complicated operation.With the increasing complexity of processing tasks,especially in the current development of personalized customization and flexible production,the current teaching methods limit the programming and use efficiency of industrial robots,and it is difficult to meet the requirements of industrial production.In general,robot teaching programming is developing towards a more intelligent and friendly interaction mode.The teaching of industrial robots based on augmented reality technology has being a hotspot of domestic and foreign scholars and industrial robot manufacturers.Taking 6-DOF industrial robots as research objects,the industrial robot augmented reality teaching programming system was proposed.Through the system,the virtual industrial robot model was superimposed on the real working scene using augmented reality technology;Using the position and attitude tracking system to collect the position and attitude of the hand-held teaching pendant and taking the pose of the pendant as input,the inverse kinematics of the robot was used to solve the pose of each joint of the virtual robot to drive the movement of the virtual robot so that the robot end-effector to reproduce path and posture of the pendant,which had achieved the purpose of the augmented reality teaching programming of the industrial robot;The Kinect physical depth camera was used to obtain the depth image of the current physical environment,and the virtual camera model was added to the virtual scene including the virtual robot model to obtain the depth image of the virtual industrial robot model;The depth image of the physical environment and the depth image of the virtual environment were used to detect collisions of the virtual robot and the physical environment,so as to verify the feasibility of the robot path planning.The main research work of the thesis was as follows:(1)An industrial robot augmented reality teaching system platform was designed and built.ProE was used to build a 6-DOF robot model and OpenSceneGraph was used to load the display for motion simulation.Using the D-H method,the kinematics of the robot was solvedand the kinematics model of the industrial robot was established.Using ARToolkit and OpenSceneGraph to build augmented reality system platform,the virtual 6-degree-of-freedom virtual robot model was superimposed on the real scene to achieve the virtual and real fusion scene so as to realise the augmented reality teaching programming.(2)A teaching method based on pose tracking system was designed.The hand-held teach pendant was designed,and the pose relationship between the handheld teach pendant and the robot end effector was established.In the augmented reality system,the attitude tracking system was designed to obtain the handheld teach pendant position data.The teach pendant pose data was used as an input to drive the virtual industrial robot motion.(3)A virtual and real collision detection algorithm based on depth image was proposed to realize the virtual and real collision detection in the virtual industrial robot teaching process.A Kinect depth camera and a virtual depth camera were added to the augmented reality system.The two depth cameras were used to acquire the depth image of the real environment and the virtual robot model respectively;and the two depth images were used to realize the virtual robot model and the real environment collision detection to optimize the path of industrial robots.(4)The related experiments were carried out for the robot augmented reality teaching system designed in this paper.The experimental results shows that the system can accurately augment the reality of the virtual robot model,and the collision detection can be avoided in real time during the teaching process.The occurrence of a collision preferably optimizes the optimal working path of the robot.Compared with traditional offline teaching,the system has better human-computer interaction and real-time performance.