It’s All in the Eyes: Designing Facial Expressions for an Interactive Robot Therapy Coach for Children

Abstract

An important aspect for child and robot interactions in various therapy scenarios is the robot's ability to convey emotions to the child. Due to the fact that 93% of human communication is non-verbal, these socially interactive robots need to have the ability to mimic non-verbal cues to the child, particularly through the use of facial expressions. In this paper, we discuss the ability for a socially interactive robot to emote emotions through a minimal set of features, i.e. soley through the eyes. In a study with five participants, we evaluate participants ability to recognize emotions based on the Plutchik emotion scale and the universal emotions of happiness, sadness, anger, and fear. Results indicate that participants recognition of an emotion is maximum when the intensity of the emotion is not at the extreme ends of the Plutchik emotion scale.