Intuitive interaction for robotic grasping

Abstract

To explore an intuitive human-robot interface for service robots, this work develops effective and user-friendly approaches which combine the real time remote vision-based teleoperation and autonomous grasping. In the first approach, hand gestures are identified and used to activate autonomy, which can reduce grasping time and provide more natural grasping poses. A potential energy algorithm is introduced to assist grasping during teleoperation in another approach. This allows us to perform online adjustments to reach a pre-grasp pose without requiring premature object contact or regrasping strategies.