Third Eye: Exploring the Affordances of Third-Person View in Telepresence Robots

Abstract

Social interaction through telepresence robots can be challenging for a robot operator due to lack of spatial awareness caused by limited idiothetic cues and narrow field of view of a robot’s camera. We explore the use of a third-person perspective, popular in video game design, to provide missing spatial cues to remote robot operators. We present the design and implementation of Third Eye, a system that enables controlling telepresence robots through a third-person view. Third Eye comprises a controllable third-person camera with a wide field of view, attached to a robot, and bimanual controls for remote operation. Observations from a user study show that Third Eye enabled the robot operators to have a better awareness of the robot ‘bodies’ they controlled. This, in turn, afforded new behavior for operators. In addition, the camera design supported ecologically valid interaction for social telepresence. Quantitative data shows that Third Eye has comparable navigation efficiency to existing systems.