My core doctoral research at Sapienza University of Rome: transforming ambient Bluetooth Low Energy radio observations into real-time, privacy-preserving spatial presence intelligence without installing invasive cameras or dedicated tracking hardware.
How passive Bluetooth Low Energy observations are transformed into meaningful occupancy, spatial, and mobility information.
High-resolution occupancy estimation in commercial and smart buildings is fundamentally constrained by privacy concerns, sensor noise, and environmental lag. This phase established the foundational mathematical framework.
Designed a novel deep learning architecture combining Graph Attention Networks (GAT) with temporal recurrence to capture spatial dependencies across adjacent zones without requiring surveillance cameras.
Rigorous empirical evaluation conducted across multi-zone academic labs, open-plan corporate offices, and conference facilities over an 18-month longitudinal study.
BLE device visibility is translated into meaningful occupancy indicators through room-specific baselines, signal filtering, and interpretable stability measures.
Processed BLE observations are associated with locations and time windows to make indoor activity patterns understandable at a glance.
The same passive signals can inform context-aware experiences by distinguishing individual and shared mobility situations without GPS or intrusive tracking.
Interactive demonstrations of infrastructure-free BLE sensing for occupancy estimation, spatial activity, and context-aware mobility.
