Digital Twin Framework for Context-Aware Safety Adherence Monitoring Using IMU-Based Digital Biomarkers

This paper presents a digital twin framework for monitoring people with dementia safety adherence using multimodal data from wearable smart glasses in home settings. Monitoring adherence to safety measures during activities of daily living (ADL) can serve as a useful proxy for evaluating cognitive and functional decline. The proposed framework integrates inertial measurement unit (IMU) data with environmental perception to infer user safety adherence behaviour relative to environmental hazards and home safety interventions using fuzzy logic (FL). The inferred behaviours are temporally aggregated to generate digital biomarkers that enable longitudinal tracking of cognitive decline. Experimental results demonstrate the feasibility of the framework in capturing and analysing safety adherence dynamics, showing its potential for long-term cognitive health monitoring.