Precision timing in human-robot interaction: coordination of head movement and utterance

As research over the last several decades has shown that non-verbal actions such as face and head movement play a crucial role in human interaction, such resources are also likely to play an important role in human-robot interaction. In developing a robotic system that employs embodied resources such as face and head movement, we cannot simply program the robot to move at random but rather we need to consider the ways these actions may be timed to specific points in the talk. This paper discusses our work in developing a museum guide robot that moves its head at interactionally significant points during its explanation of an exhibit. In order to proceed, we first examined the coordination of verbal and non-verbal actions in human guide-visitor interaction. Based on this analysis, we developed a robot that moves its head at interactionally significant points in its talk. We then conducted several experiments to examine human participant non-verbal responses to the robot's head and gaze turns. Our results show that participants are likely to display non-verbal actions, and do so with precision timing, when the robot turns its head and gaze at interactionally significant points than when the robot turns its head at not interactionally significant points. Based on these findings, we propose several suggestions for the design of a guide robot.