Comparing different metrics quantifying pedestrian safety

Abstract

Simulations of human crowds are used to improve the safety of pedestrians at events and in buildings. In this paper, we propose a methodology for comparing different metrics that are used to determine pedestrian safety. The included metrics are density, velocity, flow and pressure fields. When comparing these metrics, researchers often look at the density/velocity relationship, commonly referred to as a fundamental diagram. We propose a method for comparing these metrics without the need for fundamental diagrams or visual inspection of these fields. Furthermore, we refine existing metrics to include obstacles in these fields by replacing the Euclidean distance by the geodesic (walking) distance. In our experimental analysis we see a wide range of possible values while the flows are similar. This confirms the need for calibrating the Level of Service for each individual metric and scenario.