Joystick mapped Augmented Reality Cues for End-Effector controlled Tele-operated Robots

Abstract

End-effector control of robots using just remote camera views is difficult due to lack of perceived correspondence between the joysticks and the end-effector coordinate frame. This paper reports the positive effects of augmented reality visual cues on operator performance during end-effector controlled tele-operation using only camera views. Our solution is to overlay a color-coded coordinate system on the end-effector of the robot using AR techniques. This mapped and color-coded coordinate system is then directly mapped to similarly color-coded joysticks used for control of both position and orientation. The AR view along with mapped markings on the joystick give the user a clear notion of the effect of their joystick movements on the end-effector of the robot. All camera views display this registered dynamic overlay information on-demand. An insertion task was used to compare performance with and without the coordinate mapping using fifteen subjects. Preliminary results indicate a significant reduction in distance and reversal errors