A different approach of using Personas in human-robot interaction: Integrating Personas as computational models to modify robot companions' behaviour

The current paper focuses on a novel integration of the Personas technique into HRI studies, and the definition of a Persona-Based Computational Behaviour Model for achieving socially intelligent robot companions in living environments. Our core interest is the creation of companions adapted to users' needs to support their activities of daily living. The aim is to create a mechanism that allows us to develop initial robot behaviour, i.e. behaviour when first encountering the user, which is already adapted to each user without the necessity of collecting in advance a large dataset to train the system. A persona represents the specific needs of many individuals for a particular scenario. This technique helps us develop initial robot behaviour adapted to user needs, and so reduces the amount of trials that participants have to perform during early stages of the system development. The paper describes how this behaviour model has been created and integrated into a functional architecture, and presents the motivation, background and conceptual framework for this new research direction. Future empirical studies will validate this approach and expand the initial definition of our model.