Assisting manipulation and grasping in robot teleoperation with augmented reality visual cues

Abstract

Teleoperating industrial manipulators in co-located spaces can be challenging. Facilitating robot teleoperation by providing additional visual information about the environment and the robot afordances using augmented reality (AR), can improve task performance in manipulation and grasping. In this paper, we present two designs of augmented visual cues, that aim to enhance the visual space of the robot operator through hints about the position of the robot gripper in the workspace and in relation to the target. These visual cues aim to improve the distance perception and thus, the task performance. We evaluate both designs against a baseline in an experiment where participants teleoperate a robotic arm to perform pick-and-place tasks. Our results show performance improvements in diferent levels, refecting in objective and subjective measures with trade-ofs in terms of time, accuracy, and participants' views of teleoperation. These fndings show the potential of AR not only in teleoperation, but in understanding the human-robot workspace.