- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- Microscopic agent-based modeling and simulation of...
Open Collections
UBC Theses and Dissertations
UBC Theses and Dissertations
Microscopic agent-based modeling and simulation of cyclists on off-street paths Mohammed, Hossameldin
Abstract
Inclusion of bicycle traffic in microsimulation tools is essential for evaluating bicycle-accessible infrastructure projects. However, the representation of bicycles in microsimulation models is still at an early stage of development. A better understanding of cyclist behaviour during various interactions is needed to enhance bicycle microsimulation models, which is a pre-requisite for accurate microscopic modeling of bicycle traffic operations. Due to the limited availability of detailed data, the inherent complexity of cyclist decision-making, and the substantial heterogeneity in cycling behaviour, modeling cyclist operation behaviour requires novel methods and techniques. This thesis aims first to characterize cyclist maneuvers in following and overtaking interactions using multivariate finite mixture model-based clustering. Second, an agent-based bicycle simulation method is proposed to model cyclists as intelligent agents making operational and tactical decisions based on their observations of the operating environment. Cyclist position data associated with time stamps are used to infer state and future decisions. The data are extracted from videos collected in Vancouver, BC, Canada using computer vision techniques. For segmenting behavioural states, observations of cyclists in following interactions are clustered into constrained and unconstrained states. Observations of overtaking cyclists are clustered into initiation, merging and post-overtaking states. Generative adversarial imitation learning (GAIL) is used to infer the uncertain intentions and preferences of cyclists from observational data. The model is validated by comparing multivariate distributions of variables such as speed, direction, and spacing of observed and simulated cyclist trajectories. The model performs well in comparison to two other cyclist simulation models from the literature. The proposed approach to miscrosimulation is a significant advancement in agent-based modeling methods, with continuous, non-linear, and stochastic representation of states, decisions, and actions. By modeling cyclist heterogeneity, the proposed approach can enhance applications in bicycle facility planning and design, safety modeling, and energy modeling with consideration of the full diversity of cyclists. Such an advancement is necessary for developing bicycle networks for all ages and abilities of riders.
Item Metadata
Title |
Microscopic agent-based modeling and simulation of cyclists on off-street paths
|
Creator | |
Supervisor | |
Publisher |
University of British Columbia
|
Date Issued |
2023
|
Description |
Inclusion of bicycle traffic in microsimulation tools is essential for evaluating bicycle-accessible infrastructure projects. However, the representation of bicycles in microsimulation models is still at an early stage of development. A better understanding of cyclist behaviour during various interactions is needed to enhance bicycle microsimulation models, which is a pre-requisite for accurate microscopic modeling of bicycle traffic operations. Due to the limited availability of detailed data, the inherent complexity of cyclist decision-making, and the substantial heterogeneity in cycling behaviour, modeling cyclist operation behaviour requires novel methods and techniques. This thesis aims first to characterize cyclist maneuvers in following and overtaking interactions using multivariate finite mixture model-based clustering. Second, an agent-based bicycle simulation method is proposed to model cyclists as intelligent agents making operational and tactical decisions based on their observations of the operating environment. Cyclist position data associated with time stamps are used to infer state and future decisions. The data are extracted from videos collected in Vancouver, BC, Canada using computer vision techniques. For segmenting behavioural states, observations of cyclists in following interactions are clustered into constrained and unconstrained states. Observations of overtaking cyclists are clustered into initiation, merging and post-overtaking states. Generative adversarial imitation learning (GAIL) is used to infer the uncertain intentions and preferences of cyclists from observational data. The model is validated by comparing multivariate distributions of variables such as speed, direction, and spacing of observed and simulated cyclist trajectories. The model performs well in comparison to two other cyclist simulation models from the literature. The proposed approach to miscrosimulation is a significant advancement in agent-based modeling methods, with continuous, non-linear, and stochastic representation of states, decisions, and actions. By modeling cyclist heterogeneity, the proposed approach can enhance applications in bicycle facility planning and design, safety modeling, and energy modeling with consideration of the full diversity of cyclists. Such an advancement is necessary for developing bicycle networks for all ages and abilities of riders.
|
Genre | |
Type | |
Language |
eng
|
Date Available |
2024-01-08
|
Provider |
Vancouver : University of British Columbia Library
|
Rights |
Attribution-NonCommercial-NoDerivatives 4.0 International
|
DOI |
10.14288/1.0438571
|
URI | |
Degree | |
Program | |
Affiliation | |
Degree Grantor |
University of British Columbia
|
Graduation Date |
2024-05
|
Campus | |
Scholarly Level |
Graduate
|
Rights URI | |
Aggregated Source Repository |
DSpace
|
Item Media
Item Citations and Data
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International