Digital twin-based framework for behavioral modeling and activity recognition in smart homes

Ambient Assisted Living (AAL) systems, which combine digital healthcare technologies with Smart Home environments, play an important role in supporting independent living, especially for elderly individuals and people with physical impairments. However, the growing need to continuously monitor users, collect large and heterogeneous data streams, and anticipate abnormal or dangerous situations, has highlighted the necessity of advanced analytical frameworks capable of adapting to dynamic daily contexts.To address this necessity, this paper proposes a framework that integrates the Digital Twin (DT) paradigm with a fully implemented simulation platform. This simulator provides the required analytical capabilities by modeling user interactions with domestic objects and environmental sensors, enabling context-aware understanding of daily activities, behavioral modeling, and the generation of alternative scenarios.Experimental results show that the simulated activities closely reflect real-world patterns, producing distinctive temporal and contextual signatures that can be effectively recognized by classification algorithms. These findings underline the potential of combining DT concepts with simulation-based approaches to enhance situational awareness, adaptability, and user assistance in AAL environments.