Description
The applied research concern of preventing crowd disasters — crushes, stampedes, asphyxiation events — through better venue design, better real-time monitoring, and better operational decision-making. Crowd safety is a multi-layer defense problem: design-time simulation, deployment-time monitoring, runtime alerting, and post-incident learning. Wireless sensing contributes the runtime monitoring layer.
Why it's hard
- Real disasters are rare, fast, and catastrophic — datasets are scarce and ethically fraught.
- Safety is a system property: a safe venue depends on layout, staffing, signage, and monitoring jointly.
- Risk thresholds (density, pressure, contact rate) are venue-specific and contested in the literature.
- Operator-in-the-loop response time competes with the timescale of crowd-failure cascades.
- Liability concerns make data-sharing across operators and researchers difficult.
Common approaches
- Swiss-cheese / Vision-Zero defense-in-depth frameworks borrowed from aviation and patient safety.
- Density / pressure thresholds derived from biomechanics studies as alerting triggers.
- Crowd-entropy and other aggregated risk metrics for online decision support.
- Pre-event evacuation simulations to validate venue plans.