REHABOT: Developing an AI Assistant to Support Geriatric Rehabilitation at Home

Following discharge from geriatric rehabilitation centers (GRC), home rehabilitation offers a cost-effective pathway for older adults to achieve long-term recovery goals while alleviating capacity pressures on GR professionals. This approach emphasizes digital care, leveraging smart sensor technology and mobile health (mHealth) platforms to enable continuous monitoring and data-driven assessment and coaching. Such E-Rehabilitation systems provide GR healthcare professionals with comprehensive, objective insights into patient rehabilitation progress. However, current solutions also introduce several new clinical challenges.
Due to limited contextual understanding of patients' home environments and daily exercise experiences, GR professionals often find it difficult to effectively interpret sensor data. This gap hinders their capacity to make timely, personalized rehabilitation adjustments. Furthermore, digital feedback via E-Rehabilitation platforms lacks the nuance of in-person therapeutic interactions. This limitation hinders the delivery of clinically actionable coaching that older adults can readily comprehend and incorporate into their home exercise routines.
To bridge this gap, our project pioneers the integration of artificial intelligence (AI) in home-based E-Rehabilitation through the development of REHABOT, an intelligent rehabilitation assistive system. Specifically, we explore AI’s dual role as: (1) a AI-enabled conversational voice assistant (REHABOT Home) that engages older adults in contextual dialogues to collect rehabilitation-relevant information about their exercises, experiences, and daily activities; and (2) an AI-powered clinical analytics platform (REHABOT Insight) that supports GR healthcare professionals in data interpretation, progress assessment, and intervention generation.
A series of focus-group and co-design workshops will be organized with older adults rehabilitating at home, GR professionals, ICT specialists, AI researchers, and UX designers to explore user needs, acceptance, and technical requirements. Building on these insights, the REHABOT system, especially AI functionalities, will be developed and integrated into a sensor-based E-Rehabilitation platform. The system’s feasibility and preliminary effectiveness will then be rigorously evaluated in real-world clinical practice.