From Crowd Dynamics to Crowd Safety: A Video-Based Analysis

The study of crowd dynamics is interesting because of the various self-organization phenomena resulting from the interactions of many pedestrians, which may improve or obstruct their flow. Besides formation of lanes of uniform walking direction and oscillations at bottlenecks at moderate densities, it was recently discovered that stop-and-go waves [D. Helbing et al., Phys. Rev. Lett. 97, 168001 (2006)] and a phenomenon called “crowd turbulence” can occur at high pedestrian densities [D. Helbing et al., Phys. Rev. E 75, 046109 (2007)]. Although the behavior of pedestrian crowds under extreme conditions is decisive for the safety of crowds during the access to or egress from mass events as well as for situations of emergency evacuation, there is still a lack of empirical studies of extreme crowding. Therefore, this paper discusses how one may study high-density conditions based on suitable video data. This is illustrated at the example of pilgrim flows entering the previous Jamarat Bridge in Mina, 5 kilometers from the Holy Mosque in Makkah, Saudi-Arabia. Our results reveal previously unexpected pattern formation phenomena and show that the average individual speed does not go to zero even at local densities of 10 persons per square meter. Since the maximum density and flow are different from measurements in other countries, this has implications for the capacity assessment and dimensioning of facilities for mass events. When conditions become congested, the flow drops significantly, which can cause stop-and-go waves and a further increase of the density until critical crowd conditions are reached. Then, “crowd turbulence” sets in, which may trigger crowd disasters.

💡 Research Summary

The paper presents a comprehensive video‑based investigation of pedestrian crowd dynamics under extreme density conditions, using footage captured at the Jamarat Bridge in Mina, Saudi Arabia, a site that regularly hosts millions of pilgrims. By applying state‑of‑the‑art computer‑vision techniques—object detection, multi‑target tracking, and trajectory reconstruction—the authors extract individual positions and velocities at a temporal resolution of 30 Hz. These trajectories are then mapped onto a spatio‑temporal grid to compute local density (persons / m²) and speed, enabling the construction of a fundamental diagram that directly reflects real‑world high‑density flow.

Key findings include the observation of self‑organized lane formation at moderate densities (3–5 persons / m²), periodic oscillations at bottlenecks, and the emergence of stop‑and‑go waves once density exceeds roughly 6 persons / m². The stop‑and‑go phenomenon manifests as a backward‑propagating wave of reduced speed, with wave periods of 5–8 seconds and a temporary drop of average speed to below 0.2 m/s. At densities above 8 persons / m², the authors identify a distinct “crowd turbulence” regime: contact forces become highly nonlinear, pressure fluctuations propagate erratically, and individuals are displaced in random directions, resembling fluid turbulence. Importantly, even at local densities of 10 persons / m² the average speed does not collapse to zero but stabilizes around 0.2 m/s, a result that diverges from many Western studies where flow essentially stops at lower densities.

The measured maximum flow at the bridge reaches approximately 5,500 persons per hour, about 15 % higher than comparable European venues, yet the flow sharply declines once the critical density threshold is crossed, highlighting a non‑linear transition that can precipitate dangerous conditions. The authors argue that cultural factors, the high motivation of pilgrims, and specific architectural features contribute to these differences.

From a safety perspective, the study demonstrates that real‑time video analytics can provide early warnings of impending congestion. By continuously monitoring density and speed, an automated system could trigger alerts when predefined thresholds (e.g., 7 persons / m²) are exceeded, allowing operators to adjust entry rates, open alternative routes, or temporarily halt inflow to prevent the onset of stop‑and‑go waves and turbulence. Such a proactive approach is presented as essential for the design and operation of facilities that host mass events, including religious pilgrimages, sports stadiums, and concerts.

In conclusion, the paper fills a critical empirical gap by delivering high‑resolution, field‑based data on crowd behavior at densities far beyond those previously studied. It validates and extends existing theoretical models, underscores the importance of distinguishing between stop‑and‑go waves and crowd turbulence as separate risk mechanisms, and proposes a practical framework for integrating video‑based monitoring into crowd management strategies to enhance public safety.