Description
The application context where indoor wireless human sensing intersects with building automation: occupancy-driven HVAC, lighting, ventilation, space utilization analytics, security, and retrofitted accessibility. Smart-building sensing is broader than occupancy-aware-energy-efficiency (which is one objective) and broader than occupancy-estimation (which is one input); it is the systems-engineering frame where sensors, control loops, building management protocols, and human-comfort goals meet. The thesis's crowd-monitoring contribution is most directly deployable in this context.
Why it's hard
- Multi-stakeholder requirements: facility managers, tenants, IT, HR all want different signals.
- Legacy BMS protocols (BACnet, Modbus, KNX) constrain integration architectures.
- Per-building commissioning costs dominate — anything requiring per-room training is a hard sell.
- Long deployment lifetimes (10+ years) collide with rapid change in ML model practice.
- Privacy and labor-relations concerns make space-utilization analytics politically delicate.
Common approaches
- Multi-modal sensor fusion (CSI + BLE + CO2 + PIR + door contacts) feeding occupancy estimators.
- Standardized data layers (Brick schema, RealEstateCore) for cross-system semantics.
- Edge gateways performing privacy-preserving aggregation at the sensor.
- IoT-platform integration with cloud analytics for non-real-time use cases.
Source Papers
- chaudhari2024_6efc ↗ — occupancy-detection fundamentals for smart buildings (IoT sensors).
- khan2024_43e8 ↗ — occupancy prediction in IoT-enabled smart buildings.
- shahbazian2023_1172 ↗ — RF-signal survey of occupancy and activity detection.
- zanella2014_d33c ↗ — Internet of Things for Smart Cities (broader systems framing).