A Cognitive Approach to Model Intelligent Collaboration in Human-Robot Interaction

While Human-Robot Interaction (HRI) and Social Robotics achieved significant results, research has not given the right attention both to the role of the analysis of cognitive processes by a robot and an extensive interpretation of intelligent collaboration, i.e robot adaptivity, trust, categorization, and so on. Daily life scenarios involving robots require a kind of collaboration oriented towards the respect and the achievement of the most relevant users goals, but also to the validation and protection of other application domains elements (e.g. other agents, available resources). To address this challenge, computational cognitive models are required, in order to integrate multiple users goals and interests within the interaction context and support robots to justify the integration process. In this theoretical paper, we propose a taxonomy, achievable although with certain constraints and occasional imprecision, of the humanactionintentions into three distinct classes: practical, socio-normative, and ethical. Building upon this premise, we outline the development of computational cognitive models aimed at supporting a useful, effective, acceptable and trustworthy interaction between humans and robots. While our conceptual framework is applicable across various domains, we will illustrate it with examples drawn from the domain of Cultural Heritage