Personalized Multi-Agent Recommendation System for Monitoring and Coaching through Wearable and Non-Invasive Sensors

The aging population and the growing prevalence of chronic diseases are placing increasing demands on global healthcare systems. In this context, the remote acquisition of physiological and behavioural data through wearable and non-invasive devices plays a crucial role in supporting Active and Assisted Living (AAL) solutions. However, transforming sensor data into actionable and personalized health guidance remains a key challenge. This paper presents a novel multi-agent system (MAS) designed to support ageing individuals, healthcare professionals, and personal coaches. The system leverages real-time physiological and personal signals to deliver tailored health recommendations, support engagement, facilitate socialization, and enable proactive interventions. Each agent in the system is specialized in interpreting specific data types-such as heart rate, movement, and sleep patterns-and collaborates with others to provide a holistic understanding of the user's health status. Importantly, the behavior and decision-making of each agent are dynamically personalized based on the unique data, preferences, and evolving health conditions of each user. The system also facilitates communication between seniors, coaches, and healthcare professionals, promoting shared decision-making and continuous support. Preliminary results show the system's potential in enhancing user engagement, improving care coordination, and enabling scalable, personalized coaching in AAL settings.