Dyadic Gaze Patterns during Child-Robot Collaborative Gameplay in a Tutoring Interaction

Abstract

This study examines patterns of coordinated gaze between a child and a robot (NAO) during a card matching game, ‘Memory’. Dyadic gaze behavior like mutual gaze, gaze following and joint attention are indications both of child’s engagement with the robot and of the quality of child-robot interaction. Eighteen children interacted with a robot tutor in two settings. In the first setting, the robot tutor gave clues to assist children in finding the matching cards, and in the other setting, the robot tutor only looked at the participants during the play. We investigated the coordination between child and robots’ gaze behaviors. We found that more occurrences of mutual gaze and gaze following made the children aware of the gaze hints given by the robot and improved the efficacy of the robot tutor as a helping agent. This study, therefore, provides guidelines for gaze behaviors design to enrich child-robot interaction in a tutoring context.