A Semantic Framework to Detect Problems in Activities of Daily Living Monitored through Smart Home Sensors

Activities of daily living (ADLs) are fundamental routine tasks that the majority of physically and mentally healthy people can independently execute. In this paper, we present a Semantic Framework for detecting problems in ADLs execution, monitored through Smart Home sensors. In the context of this work, we conducted a pilot study, gathering raw data from various sensors and devices installed in a smart home environment. The proposed Framework combines multiple Semantic Web technologies (i.e. ontology, RDF, triplestore) to handle and transform these raw data into meaningful representations forming a Knowledge Graph. Subsequently, SPARQL queries are used to define and construct explicit rules to detect problematic behaviours in ADLs execution, a procedure that leads to generating new implicit knowledge. Finally, all available results are visualized in a clinician dashboard. The proposed framework can monitor the deterioration of ADLs performance for people in the spectrum of dementia, by offering a comprehensive way for clinicians to describe problematic behaviors in the every-day life of an individual.