A study on the people's movement in normal and stressful conditions

Abstract

The past several decades have witnessed the increasing interests in the study of pedestrian movement modeling and crowd dynamics. Getting insights into pedestrian movement modeling is important for public space designers, dealing with space constraints and congestion problems, and for transportation researchers, dealing with facilities integration, demand modeling and safety issues. Many comprehensive theoretical models have been developed and some of them even became commercial applications. The main research objective of this thesis is to develop a dedicated framework for studying the people's movement in normal and stressful conditions. The developed framework could support building designer and facility planners to optimize their designs, making it safer, more efficient and more comfortable. In order to achieve the main objective, a series of research works have been conducted. In the literature review, an in-depth account of existing knowledge on pedestrian movement related studies as well as an overview of the blank spots was given. Especially, an intensive overview of empirical studies on pedestrian flow was presented, which was divided into ten categories in terms of the research emphasis. The review almost covered every single aspect about empirical studies, and could offer a great reference for both academic and engineering use. Two major research approaches, pedestrian walking behavioral modeling and pedestrian trajectory tracking, were proposed. Instead of developing a "new pedestrian behavioral model, focus was concentrated to improve the existing model, making it better applied to analyze real-world pedestrian flow scenario. An elaborated pedestrian behavioral model by enhancing the Social Force model was established in this context. The enhanced model was equipped with more elaborated mathematical structures and was better designed for practical use. To provide a general means which was capable of collecting pedestrian trajectories from video recordings in almost all circumstances, not just constrained to, for instance, a top view or a good lighting condition, an improved mechanism of extracting pedestrian walking trajectories from video recordings was developed. The pedestrians were detected manually, and recursively trajectory prediction was performed, using Extended Kalman Filter (EKF) approach to enhance the tracking process. This pedestrian tracking method could benefit the calibration and validation of the developed microscopic pedestrian behavioral models, and more importantly, could possibly discover the underlying theory embedded in the empirical data. The first applications works, empirical studies for vertical facilities, were reported. The review of previous works indicated that there were inadequate empirical works having been done to support the effective evaluation of vertical facility utilization. With the developed pedestrian trajectory tracking tool, empirical studies were carried out to determine a couple of pedestrian flow characteristics on the use of vertical facilities. The empirical results could be incorporated into the established microscopic pedestrian behavioral model to simulate the utilization of vertical facilities for varying traffic demands. Finally, two case studies associated with the refurbishment in a mass transit station were conducted. The first study was to investigate the pedestrian traffic conditions in the platform and to find the appropriate measures to improve the traffic efficiency. The second study was to determine the people's pre-movement time in the concerned station and to reveal the impact of the pre-movement behavior on the evacuation process.

Date of Award	15 Feb 2011
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Siu Ming LO (Supervisor)