Designing Synthetic Emotions of a Robotic System

As massive changes in our lifestyle and society structure are coming, due to the contemporary pandemic situation, we face more than ever the challenge of employing non-human and teleassistance devices in many areas, especially those that require assisting and caring for weak users. Humanoid robots are proven to be a valuable asset in these situations but require to be carefully designed in their interaction features, in order to be accepted and valuably used by their users. In particular, this study is focused on robots that have a certain degree of human-likeness that allows to define them “humanoids”; for this kind of robots we can say that the area devoted to replicate human facial features is the most important interface for human–robot interaction. Actually, more than 60% of human–human interaction is conducted non-verbally, by using facial expressions and gestures. For a robot to be able to engage in this kind of interaction with a human and provide understandable feedbacks is a massive step towards acceptance and development of an affectional relationship by users. Being meant to reproduce human emotions, visual feedbacks are mostly developed referring to eminent researches in the psychological field: as Paul Ekman already observed in 1998, human faces have a universal coding for six basic expressions that represent as many basic emotions: fear, anger, disgust, happiness, sadness and surprise. Our research focuses on the design of an expression system to be implemented in a European-funded project for an assistive robot that will support weak users at home or in assistance facilities and their caregivers. Our main concern regarding this project are to design a dynamic, human-friendly system that is scalable and visually recalls real facial expressions without being too much human-like, in order to avoid the well-known uncanny effect described by Masairo Mori.