Conceptual Framework for Autonomous Coaching in Orthopaedic Rehabilitation with Socially Assistive Robots

Post-surgical orthopaedic rehabilitation requires personalized and adaptive therapy to restore motor function, yet current practices face challenges related to patient variability, diverse supervision needs, and limited treatment personalization. Socially Assistive Robots (SARs) offer promising support, but existing systems often rely on predefined exercise scripts and heuristic adaptation, with limited autonomy. In this paper, we propose the use of a cognitive framework to enable SARs to autonomously coach orthopaedic rehabilitation sessions through the integration of automated planning based on structured clinical knowledge with multimodal monitoring. By combining symbolic and numerical planning with closed-loop feedback, the system can adapt rehabilitation sessions in response to the patient’s performance and provide feedback.