Investigating biomechanical and spatial dynamics of pedestrian movement in the context of microscopic crowd flow modelling

Abstract

Crowd movement models are largely based on historical data and include various assumptions, which could potentially lead to inaccuracies. This thesis aims to leverage experimental data to explore the movement parameters of individuals in a crowd and their inter-relationships, to inform the development of a mathematical crowd movement model. The development of more detailed and accurate models enhances our understanding of crowd behaviour for building design and crowd management strategies, potentially improving safety and level-of-service in congested spaces.

The first phase of this study involved an in-depth review of literature across various fields, including fire engineering, transportation engineering, and biomechanics, to identify gaps and provide a foundation for further research in crowd dynamics. The second phase involved the extraction and detailed analyses of important movement parameters and their interrelationships from video footage and motion capture devices used in single-file and wider flow experiments designed to support the development of the movement adaption model. Specifically, it explores the inter-relationships between gait characteristics, i.e., step length, step extent, and distances, i.e., contact buffer and inter-person distance of individuals moving in a crowd, with speed. The core equations of the model are also explored and confirmed. Importantly, a novel approach to the characterisation and quantification of the stop/start walking process of individuals in a congested space is presented. The components of the stop/start walking, i.e., perception-reaction time, slow-down time, and start-up time delay were0.48 s, 0.58 s, and 0.39 s, respectively. Finally, the findings from the in-depth experimental analyses are used to inform/modify the movement adaption model for single-file movement. The results of the updated model are compared with the results of other experimental single file studies, and alternative approaches to the further development of the model are presented. Additionally, the outcomes of these analyses are summarised in a form which may be used to inform crowd movement models.

Date of Award	Mar 2024
Original language	English
Supervisor	Nigel Mc Connell (Supervisor), Jianping Zhang (Supervisor) & Karen Boyce (Supervisor)