Empirical Analysis of Pedestrian Merging Process with Different Merging Angles and Merging Layouts

Hanchen Yu,Nan Jiang,Hongyun Yang,Jixin Shi,Zhenyu Han,Eric Wai Ming Lee,Lizhong Yang
DOI: https://doi.org/10.1016/j.physa.2024.130218
IF: 3.778
2024-01-01
Physica A Statistical Mechanics and its Applications
Abstract:Crowd gathering in public infrastructures increases the challenges to safety management. To control the risk of crowd accidents, it is crucial to investigate the specific pedestrian movement characteristics within typical facilities to ensure efficient and safe walking operations. Merging scenarios involve complex multi-directional movements and always be considered dangerous bottlenecks to trigger serious crowd accidents. It demonstrated that the merging angle and layout significantly affect the performance of merging structures. However, most studies treat those properties independently without systematically exploring their combined effects to reflect a more diverse and complex flow status in real situations. In this study, we conduct controlled experiments to investigate the pedestrian merging process with five different merging angles and three different merging layouts. By analyzing the movement status, congestion, and degree of order, the efficiency and risk of different merging scenarios are compared. The lower upstream velocity and faster clogging propagation can be observed as the merging angle increases. While in scenarios with a small merging angle under certain layouts, limited outflows and longer time intervals occur and the stagnation is more frequent. The walking space of pedestrians near the inner side is severely compressed, resulting in higher local density around the corner and exacerbating the clogging nearby. Furthermore, a new indicator named Harmony Index is proposed to describe the adaptability of crowds to different merging setups. These findings are beneficial for optimizing the safe design and crowd management strategies of public infrastructures, as well as the crowd movement setup in relevant modeling works.
What problem does this paper attempt to address?