Dynamic analysis of pedestrian movement in single-file experiment under limited visibility

• Published in: Communications in nonlinear science & numerical simulation 2019-04, Vol.69, p.329-342
• Main Authors: Cao, Shuchao, Wang, Peng, Yao, Ming, Song, Weiguo
• Format: Article
• Language: English
• Subjects: Algorithms; Diagrams; Flow-density-speed relationships; Fundamental diagrams; Light transmission; Measurement methods; Normal distribution; Pedestrian dynamics; Pedestrian safety; Pedestrian traffic flow; Pedestrians; Single-file experiment; Velocity; Velocity distribution; Visibility
• ISSN: 1007-5704; 1878-7274
• DOI: 10.1016/j.cnsns.2018.10.007

•Pedestrian dynamics under different visibilities is studied by experiments.•Two regimes are found in the headway–velocity relation under limited visibility.•Fundamental diagrams of pedestrian flow under different visibilities are obtained. In this paper, the movement properties of pedestrians under limited visibility are investigated by a group of single-file experiments. Three light transmissions (0.3%, 0.1% and 0.0%) are set and trajectories of individuals are extracted by using tracking algorithm. From the analysis of experimental video, typical behavior of following the boundary and frontal pedestrian are observed under poor visibility. Velocity distribution in different experiment runs conforms to Gaussian distribution. Two regimes are found in the headway–velocity relation and the time-space diagram reveals the evolution of pedestrian traffic with the decrease of visibility. Fundamental diagrams are obtained through the proposed measurement method and three regions can be found in the density-flow relation. The maximum specific flow under the light transmission of 0.3%, 0.1% and 0.0% is 1.3 s−1, 1.15 s−1 and 0.9 s−1 respectively. Finally log transformation for density and velocity is performed to study the elastic effect of density change on velocity change. It is found that pedestrians are more sensitive to the density change under good visibility. This study is helpful to understand pedestrian movement characteristics under different visibilities, which is of practical importance for pedestrian safety under emergency condition. Furthermore, the valuable data obtained from the experiment can be used for model validation and verification in the future.