Adult–child pairs walking down stairs: empirical analysis and optimal-step-based modeling of a complex pedestrian flow, with an exploration of flow-improvement strategies

Abstract

Pedestrian egress from training schools in the after-class period (especially in China, as children walk down stairs together with their parents) raises practical concerns related to degraded flow conditions and possible safety hazards. But it also represents a challenge to mainstream modeling approaches for several reasons: they involve heterogeneous groups (adult–child pairs), which are hardly studied compared to their homogeneous counterparts, in a complex geometry, made of staircases connected by a platform where pedestrians rotate, and over a wide range of densities. In light of our field observations at a training school in China, we develop a semi-continuous model which quantitatively reproduces the collective dynamics observed empirically and enables us to assess some guidance strategies to improve egress efficiency. In this model, which extends the optimal step approach, adults and children are described as ellipses with prescribed relative positions that evolve by successive steps handled by an event-driven algorithm, along a spontaneous semi-circular path on the platform, but may deviate from it in crowded conditions by selecting their preferred next positions. In line with the observations, most pairs walk side by side overall and only a small fraction sometimes make a detour to avoid queuing. Turning to the guidance strategies, we find that promoting front-back pairing may increase the flow, while urging people to make more detours may be counterproductive. Perhaps even more relevantly, the intuitive measure consisting in desynchronizing the flows from the different floors by shifting the ends of classes succeeds in reducing the egress time by at least 10%.