Description

Identifying which individuals have been in close-enough proximity for long-enough to constitute an epidemiologically significant contact. The COVID-19 era turned this from a public-health niche into a major wireless-sensing application. Two major architectures dominate: BLE-based decentralized exposure notification (GAEN-style), and infrastructure-based proximity tracking that consumes BLE / WiFi observations from fixed sensors plus an indoor map. The latter framing connects directly to indoor-localization and occupancy-estimation.

Why it's hard

  • Distance estimation from BLE RSSI is noisy and direction-of-arrival-dependent.
  • "Significant contact" is an epidemiological judgement, not a fixed distance/duration threshold.
  • Privacy-vs-utility tension is severe; centralized infrastructure approaches are politically fraught.
  • Multi-modal fusion across BLE, WiFi probe-requests, and indoor maps produces brittle systems.
  • Cross-jurisdiction interoperability is essentially absent.

Common approaches

  • BLE peer-to-peer decentralized exposure notification (Apple/Google framework).
  • Infrastructure-based BLE/WiFi capture combined with indoor-map proximity reasoning.
  • IndoorGML-grounded person-to-person and person-to-place tracking.
  • RSSI-to-distance calibration plus duration thresholds for contact event definition.

Source Papers

  • ojagh2020_321f — IndoorGML-based COVID-19 person-to-person and person-to-place contact tracing.
  • wang2019_d6f9 — survey on human-behavior recognition using CSI (related sensing).

3 vault papers address this problem

Titles and DOIs only — no abstracts, no analyses.

  • Towards Energy Efficient Wireless Sensing by Leveraging Ambient Wi-Fi Traffic 2024 DOI ↗
  • Human Sensing by Using Radio Frequency Signals: A Survey on Occupancy and Activity Detection 2023 DOI ↗
  • A Person-to-Person and Person-to-Place COVID-19 Contact Tracing System Based on OGC IndoorGML 2020 DOI ↗