Preliminary personality model for social robots based on the Cognitive-Affective Processing System theory

Abstract

Personality influences human behaviour during social interactions as well as is influenced by the situation in which these interactions take place. To date, the reciprocal interplay between personality and the situation is still little investigated in modelling personality for social robots, even if the personality psychology field has widely recognised this interplay. Therefore, in this study, we propose a novel personality model for social robots based on the formalisation of the Cognitive-Affective Processing System theory of personality. According to this theory, we conceived a personality model as a system that interacts with features of the situation to generate the behaviour. We focus on a bottom-up approach by detailing the core components of the system and by posing particular emphasis on its link with induced emotions. Furthermore, we preliminary implemented the proposed personality model taking advantage of the flexibility offered by expert systems and endowing the personality model with some degree of explainability.Our work approaches social robot personality focusing on increasing the human-likeness of the model, rather than improving the imitation of human behaviours, to improve the believability and the user experience of social human-robot interactions.