An Architecture for Integrating Social Interaction in Upper-Limb Rehabilitation

Robot-assisted physical rehabilitation holds the potential to increase the efficiency and accessibility of therapy by assisting therapists in delivering
consistent training over extended periods and reducing the workload on human therapists. In this work, we propose a system architecture that integrates social interaction into existing robot-assisted physical rehabilitation frameworks with the potential to enhance user comprehension and engagement. The theoretical basis for designing the proposed architecture is based on previous design recommendations for single-robot systems and expanded to multi-robot systems. We provide an example of how to implement the system architecture with root systems where investigate different techniques to generate dynamic speech and gestures for the socially assistive robot, which helps reduce monotony in the user experience. Furthermore, a pre-participation evaluation identified Interest-Enjoyment as the most effective motivational dimension that Pepper could potentially enhance compared to other dimensions. The architectural proposal is expected to serve as a foundation for further implementation and development of socially interactive rehabilitation systems to create a more interactive therapeutic experience that could contribute to current rehabilitation practices.