基于数字孪生的遥操作机器人在线示教系统

To improve the teaching efficiency and safety of teleoperation robot, a digital twin framework for teleoperation robot was proposed. To map physical entity to the digital twin model, the virtual models of both robot and its working scene were established. A teleoperation method based on RGB-D image and posture teaching device was designed and implemented. In this method, the paths and postures of the virtual robot end-effector were controlled by mouse and posture teaching device. In addition, the Augmented Reality (AR) technology was used to combine the virtual modes with physical scene, and the AR-based teaching of teleoperation robot was realized. The octree models of the virtual robot and its working scene were also constructed to predict the collision interference between the virtual robot and the working scene. By collision detection technology, the system sent the planned paths and postures without collision interference to the physical robot controller to control the movement of the physical robot, so as to realize the robot online teaching of the teleoperation robot.