Activity Recognition and Personalized Feedback Solution for Active and Healthy Ageing (original) (raw)

As the average lifespan increases, the care of diseases related to lifestyle and age, such as chronic and neurodegenerative ones, becomes costlier and less accessible, highlighting the need for self-management using technology. This paper proposes a pervasive system for autonomous healthy ageing, which integrates two layers of intelligence via semantic interpretation: multi-modal sensor fusion from smart devices, wearables and multimedia, and personalized spoken feedback based on context-sensing and user input. Aiming for a practical, acceptable system, the proposed architecture considers aspects of integration, security, privacy and cost. The currently implemented components include activity recognition and problem detection, complemented by end-user applications and personalized spoken feedback. The proof-of-concept implementation is evaluated both in a lab setting, for the more complex personalized feedback component, and in four real home environments, presenting efficient activity recognition, and improvement in several neuropsychological areas, such as mood, physical functional and cognitive condition of elders.