The CARE project aims to explore and develop new assistive features for a companion robot capable of monitoring and assessing the physiological and functional parameters of elderly people with multimorbidity and frailty. The project involves an initial phase of systematic literature analysis to identify the most effective technological solutions for the assessment and treatment of the main syndromes associated with ageing. The results of this research will guide the selection and integration of low-cost wearable and non-wearable sensors capable of collecting health data in real time. Subsequently, the project will focus on the identification, study and design of advanced artificial intelligence models, such as conversational agents, classifiers and multi-sensor fusion systems, to deduce the patient's cognitive state. Finally, the research group will focus on the functional design of the CARE robot prototype, which will use advanced LLM models to actively contribute, in synergy with the data collected by the sensors, to monitoring the patient's health status (both cognitive and physical) and to generate information on the patient's current and over time health status. the system will be able to provide empathetic recommendations and proactive responses even in the event of an emergency. The CARE project will not only be relevant to the health of the elderly, but could also be used as a model in the future to address issues in other areas, such as telemedicine and home care for the chronically ill.
In the first SAL, the CARE project completed the activities planned in WP1 and WP2, laying the methodological and technological foundations for the assistive robot prototype. In WP1, technical and scientific coordination was activated, with a project committee, communication protocols, activity monitoring and internal validation of results. Dissemination activities were launched with workshops, informative seminars and publications accepted at the IAS-19 and Human-friendly Robotics 2025 conferences, in addition to the creation of the project website (Milestone M3). In WP2, the analysis deliverables (M3) were completed: systematic review of digital technologies for active ageing, mapping and evaluation of wearable and non-wearable sensors (biometric, environmental, motion, computer vision), with selection of the most compatible for integration into CARE. In WP3, framework design activities began: preliminary development of multimodal data fusion models (audio-video, EEG), LLM-based pipeline for cognitive exercises (memory, attention, executive functions) and definition of the architectural mockup of the CARE system, including the sensor layer, AI and robotic social interaction. Progress is in line with the schedule, with no delays or significant issues.