Issue link: https://htpgraphics.uberflip.com/i/1385717
UNIQUENESS // To achieve collision-free operation in telemanipulation, we have been developing a novel shared-control inverse kinematics method, called Collision-Free IK, by which any part of the robot manipulator can avoid known obstacles within its operating space. On the contrary, most existing methods can only address the collision avoidance of the robot end-effector, not the entire body. Our work's key idea is to use a concept of virtual safety capsules, which encapsulate the known obstacles and encode the collision avoidance constraints into the inverse kinematics being used. The Collision-Free IK only focuses on preventing any possible collision, not providing a collision-free trajectory to a target position as that is similar to what the other existing methods do. This shared autonomy feature is favorable for real-time telemanipulation as it has less computation time for collision avoidance, and a human operator takes the role of trajectory generation anyway. LEAD RESEARCHER: INMO JANG I am Inmo Jang, a Principal Robotics Research Scientist/Engineer at Samsung and an Honorary Research Associate at the University of Manchester. In the RAIN Hub, I have been working with the Human Robot Interaction Group, particularly investigating what would be the best experience for teleoperators by introducing virtual reality-based immersive user interfaces and a shared autonomy-based collision-free inverse kinematics. Although I am now based in S Korea, thanks to the experience I gained within RAIN, I still work in robotic inspection and maintenance, but now, for the semiconductor manufacturing industry 's infrastructure facilities, which also has risky environments. SUMMARY // For a glovebox telemanipulation robotic system, the robotic arm must not cause any damage to objects inside the glovebox or to the glovebox itself and without putting containment at risk. Without this preventive capability, using a robotic manipulator for glovebox operations may fall into a regulatory issue. Furthermore, a user may encounter large cognitive load or frustration because failures cannot be allowed during operations. Substantial training is crucial, but still, there will be a psychological gap between performing mock-up tests and actual tasks. Hence, this work aims to enable a robotic manipulator to autonomously prevent any accidental collisions while in teleoperation by using a shared-autonomy framework. SHARED AUTONOMY FOR COLLISION-AVOIDING TELEMANIPULATION "WE IMPLEMENTED THE COLLISION-FREE IK ON ROS AND DEVELOPED IT TO BE USED WITH AS MANY COMMERCIAL ROBOT MANIPULATORS AS POSSIBLE SO THAT ANY 6-DOF OR 7-DOF ROBOT SUPPORTING ROS CAN BE EASILY REPLACEABLE FOR A TELEMANIPULATION SYSTEM." 20