Crowd dynamics at turning phenomena: experiments and modelling

by Charitha Gayan Dias

Institution: Monash University
Department: Department of Civil Engineering
Year: 2015
Keywords: Pedestrian crowd dynamics; Crowd modelling; Human behaviour modelling; Pedestrian simulation
Record ID: 1050020
Full text PDF: http://arrow.monash.edu.au/hdl/1959.1/1145039


With the rapid urbanization and the advancement of sustainable transport modes, major public infrastructures are increasingly visited and utilised by large crowds for daily commuter purposes and during special events. Interactions between pedestrian crowds and complex geometrical and architectural features at those public spaces can considerably impede collective movement of crowds. Therefore, it is important to understand pedestrian walking characteristics associated with common complex geometrical settings in order to plan and design these mass gathering places for day-to-day activities as well as for emergency evacuations. This is essential to ensure safety and efficiency of crowd dynamics by optimizing architectural designs and proper management of crowds at public buildings and built environment. Currently, pedestrian crowd simulation tools are widely used in designing built environments and crowd management. Those tools are useful not only in detecting problematic places or dangerous spots at crowd gathering places, but also in assessing design solutions for those places. In order to increase the reliability of such tools, they must be calibrated against reliable empirical data for a variety of situations. Detailed empirical studies could serve this purpose, while providing more insights into microscopic pedestrian flow characteristics under different conditions. As revealed through a comprehensive literature review, a major gap in the knowledge is that no substantial research has examined the impacts of complex geometrical settings, particularly turning configurations, on pedestrian crowd movements under different conditions. This dissertation investigates the impact of turning configurations and various speed and density conditions on pedestrian crowd dynamics. A wider range of turning angles is considered in this study. Comprehensive empirical investigations are conducted with the data collected through experiments with ants (under panic conditions), experiments with humans under three different desired speed levels, i.e. normal speed (1.44 m/s), fast speed (2.00 m/s) and slow running (3.00 m/s), and three different density levels, i.e. free-flow or solo, medium density (0.75 peds/m2) and high density (2.2 peds/m2). The series of experiments presented in this study is an original approach to investigate the effect of turning configurations on collective pedestrian walking characteristics. The trajectory data extracted from human experiments are utilized not only for empirical analyses, but also to identify and derive behavioural characteristics that are specific to walking through angled or turning configurations. Based on the behavioural characteristics derived through these experiments, a microscopic force-based modelling framework is also developed in this study. The deceleration effect and the centrifugal effect associated with turning configurations are considered in this modelling framework. The proposed modelling framework is calibrated and verified for different situations (for different desired…