Facial gesture is an important social cue which is subconsciously expressed in human-human interaction. It is particularly fascinating for human-robot interaction as previous study revealed that social cues affected people’s engagement in HRI. A dialogue-based interaction was designed using social robot Furhat to investigate whether adding different facial expressions could influence a human to move in the interaction space. Furhat was given two personality traits, a positive and a negative persona with relevant facial gestures. The results showed a higher degree of upper body movements across participants when interacting with the negative robot. A higher level of engagement and emotional values were found in the negative setting. The positive persona in combination with the facial gestures was perceived more unnatural, even though the overall impression was rated more natural and pleasant compared to the negative persona. Furthermore, the female participants showed sympathy towards the robot in the positive setting.

Contribution

This is a group project cooperated with 3 other KTH students. In this group, I contributed to the dialogue as well as the facial expression programming of Furhat through Furhat SDK and Kotlin. I also worked as the Wizard to control Furhat's interaction towards participants during the experiments, and handled with data resulted from the experiments using Python.