Gast ::
| Hauptseite > Publikationsdatenbank > Single-file Movement: Literature Review, Empirical Analysis with Artificial Neural Networks, and Modeling |
| Hauptseite > Publikationsdatenbank > Single-file Movement: Literature Review, Empirical Analysis with Artificial Neural Networks, and Modeling |
| Book/Dissertation / PhD Thesis | FZJ-2025-03883 |
2025
Forschungszentrum Jülich GmbH Zentralbibliothek, Verlag
Jülich
ISBN: 978-3-95806-843-8
This record in other databases:
Please use a persistent id in citations: doi:10.34734/FZJ-2025-03883
Abstract: Several influential factors impact pedestrian movement within crowds, making their analysis complex and challenging. To address this, we employ a simplified system referred to as single-file. In this system, pedestrians walk along a narrow path without overtaking, ensuring that the order of individuals remains constant. This setup reduces the number of variables and allows for a focused examination of the specific factors researchers aim to investigate in pedestrian dynamics. Given the significance of single-file movement in understanding complex movement behaviors, this thesis demonstrates the importance of studying single-file systems. Furthermore, this thesis analyzes the interaction ranges in single-file systems by incorporating into the speed model the influence of both the pedestrian ahead and the one behind, taking into account their respective distances and speeds. This novel approach, detailed in Publications II and III, enhances the accuracy of modeling in single-file movement. This cumulative thesis comprises three publications aimed at investigating pedestrians’ single-file movement. Publication I provides a comprehensive review of experiments on single-file pedestrian movement, emphasizing its importance. The review covers the historical background of single-file movement studies and offers insights from human and non-human traffic systems. The publication also elaborates on various experimental setups and data collection methods and discusses factors influencing pedestrian movement. Additionally, the study introduces a new Python-based tool, SingleFileMovementAnalysis, designed to analyze the data of pedestrian movement, particularly head trajectories, which helps prepare and calculate movement quantities such as speed, density, and headway. The publication offers an approach to experimental data analysis and suggests future directions for research in this field. In Publication II, the factors influencing pedestrian movement in single-file experiments are explored. Feed-forward neural networks are utilized to predict individual pedestrians’ speeds, using various combinations of distances and interaction ranges with neighboring pedestrians. Therefore, the influence of introducing the distance behind into the speed model is analyzed, and the predicted individual speeds using different influential factors are evaluated and compared. Inspired by the results from the statistical investigations conducted in PublicationII, Publication III introduces a new microscopic speed model that considers the relative distances to the nearest neighbors both behind and ahead for single-file movement. A fine-tuning of the weighted asymmetry parameters is applied, and the stability of the new model is analyzed. Furthermore, a numerical simulation of one-dimensional movement evaluates the proposed model.
; 
|
The record appears in these collections: |
PDF