The Role of Visual Perspective Taking in Human-Robot Collaboration

We present a video demonstration of our human-robot collaboration experiment investigating visual perspective taking and trust. In the experiment, participants interact with a Vector robot to help it navigate out of a maze. Both robot and participant have a unique view of the maze, with participants viewing it from above and Vector from a point-of-view perspective. The maze itself includes a series of junctions with several route options. After announcing the task objectives, the robot navigates autonomously and uses verbal utterances to communicate route options and route suggestions at each junction. Junction decisions are made after Vector announces the route options, listens to participants' selection, offers a route suggestion, and processes participants' final decision. So, to efficiently navigate to the mazes exit, participants' must assume Vector's visual perspective to understand the direction it is referring to. As outlined in the conclusion, this experiment intended to investigate the role of visual perspective taking on trust dynamics in collaborative human-robot interaction, with a view to supporting further work on robot theory of mind (ToM).