Date of Defense

18-4-2025 8:00 AM

Location

F1 - 1117

Document Type

Thesis Defense

Degree Name

Master of Science in Civil Engineering (MSCE)

College

COE

Department

Civil and Environmental Engineering

First Advisor

Dr. Hamad Al Jassmi

Keywords

Road Safety, Eco-Mobility, Bayesian Belief Network (BBN), Cost-Benefit Analyses (CBA), Policy Recommendation, Decision-making, Machine Learning, Crash Severity Assessment, Deep Learning

Abstract

This study presents a novel approach to prioritizing road safety mitigation measures for eco-mobility modes in Abu Dhabi, employing a sophisticated Bayesian methodology integrated with cost-benefit analysis. The research addresses the pressing need for enhanced safety in sustainable urban transportation, focusing on pedestrians, cyclists, and users of micro-mobility devices. A comprehensive Bayesian Belief Network (BBN) model was developed, incorporating key variables influencing eco-mobility safety, including visibility, predictability, road curvature, obstructions, and lighting conditions. The model was constructed using a combination of historical accident data and Probabilistic reasoning. This approach allowed for the quantification of complex relationships between various risk factors and their impact on accident occurrence. The study analyzed data from 2020 to 2023, encompassing a wide range of accident types and causes specific to eco-mobility modes. The BBN model revealed that high visibility coupled with high predictability could reduce accident probability by up to 90%. Road curvature emerged as a critical factor, with sharp curves reducing high visibility conditions from 70% to 10%. Major obstructions were found to decrease high predictability conditions from 80% to 30%. In addition, the study evaluates Bayesian Belief network (Supervised BBN with Naïve Bayes), traditional machine learning (Logistic Regression, Random Forest, Gradient Boosting, AdaBoost, SVM, KNN, Decision Tree, Naive Bayes, Extra Trees, and Bagging) and deep learning (Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) models for crash severity assessment. Among the models tested, BBN emerged as the best-performing model, with the overall precision (0.9991), mean precision (0.9992), R2 (0.9985), highlighting its potential for improving crash severity assessments in future safety initiatives. Integrating the BBN with cost-benefit analysis, the research prioritized safety interventions based on their potential impact and economic efficiency. Traffic calming measures in residential areas and school zones were identified as high-priority interventions, with a predicted 15% reduction in accidents. Increased traffic police presence and targeted enforcement campaigns showed a 12.5% potential reduction in accidents. The study also explored the implications of these findings for urban planning and policymaking in Abu Dhabi. Recommendations include both shortterm actions for immediate impact, such as enhanced visibility measures and rapid implementation of traffic calming devices, and long-term strategies for sustainable improvement, including comprehensive infrastructure overhauls and the integration of smart city technologies. Limitations of the study, including data constraints and generalizability considerations, were acknowledged, and future research directions were proposed. Overall, this research contributes significantly to the field of eco-mobility safety by providing a robust, data-driven framework for decision-making. The methodology developed here offers a template for other cities seeking to enhance the safety of sustainable transportation modes. The findings and recommendations have the potential to transform Abu Dhabi's approach to eco-mobility safety, paving the way for a more sustainable and secure urban transportation system.

Share

COinS
 
Apr 18th, 8:00 AM

BAYESIAN BELIEF NETWORKS APPROACH TO PRIORITIZE ROAD SAFETY MITIGATION MEASURES FOR ECO-MOBILITY MODES

F1 - 1117

This study presents a novel approach to prioritizing road safety mitigation measures for eco-mobility modes in Abu Dhabi, employing a sophisticated Bayesian methodology integrated with cost-benefit analysis. The research addresses the pressing need for enhanced safety in sustainable urban transportation, focusing on pedestrians, cyclists, and users of micro-mobility devices. A comprehensive Bayesian Belief Network (BBN) model was developed, incorporating key variables influencing eco-mobility safety, including visibility, predictability, road curvature, obstructions, and lighting conditions. The model was constructed using a combination of historical accident data and Probabilistic reasoning. This approach allowed for the quantification of complex relationships between various risk factors and their impact on accident occurrence. The study analyzed data from 2020 to 2023, encompassing a wide range of accident types and causes specific to eco-mobility modes. The BBN model revealed that high visibility coupled with high predictability could reduce accident probability by up to 90%. Road curvature emerged as a critical factor, with sharp curves reducing high visibility conditions from 70% to 10%. Major obstructions were found to decrease high predictability conditions from 80% to 30%. In addition, the study evaluates Bayesian Belief network (Supervised BBN with Naïve Bayes), traditional machine learning (Logistic Regression, Random Forest, Gradient Boosting, AdaBoost, SVM, KNN, Decision Tree, Naive Bayes, Extra Trees, and Bagging) and deep learning (Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM) models for crash severity assessment. Among the models tested, BBN emerged as the best-performing model, with the overall precision (0.9991), mean precision (0.9992), R2 (0.9985), highlighting its potential for improving crash severity assessments in future safety initiatives. Integrating the BBN with cost-benefit analysis, the research prioritized safety interventions based on their potential impact and economic efficiency. Traffic calming measures in residential areas and school zones were identified as high-priority interventions, with a predicted 15% reduction in accidents. Increased traffic police presence and targeted enforcement campaigns showed a 12.5% potential reduction in accidents. The study also explored the implications of these findings for urban planning and policymaking in Abu Dhabi. Recommendations include both shortterm actions for immediate impact, such as enhanced visibility measures and rapid implementation of traffic calming devices, and long-term strategies for sustainable improvement, including comprehensive infrastructure overhauls and the integration of smart city technologies. Limitations of the study, including data constraints and generalizability considerations, were acknowledged, and future research directions were proposed. Overall, this research contributes significantly to the field of eco-mobility safety by providing a robust, data-driven framework for decision-making. The methodology developed here offers a template for other cities seeking to enhance the safety of sustainable transportation modes. The findings and recommendations have the potential to transform Abu Dhabi's approach to eco-mobility safety, paving the way for a more sustainable and secure urban transportation system.