Date of Award
May 2024
Degree Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Engineering
First Advisor
Xiao Qin
Committee Members
Habib Tabatabai, Jun Zhang, Yin Wang, Chao Zhu
Keywords
extreme value theory, safet routing, surrogate safety measures, traffic conflict, traffic safety
Abstract
This dissertation demonstrates a comprehensive analysis, evaluation, and application of safety surrogate measures. Through five distinct analyses, the research has made several key contributions.Firstly, the research developed a complete framework to process and analyze the large-scale safety Pilot Model Deployment (SPMD) connected vehicle data, allowing the extraction of meaningful safety insights. By building surrogate safety measures on the vehicle, trip, and link levels, the analyses have been capable of statistically relate those measures to real crash facts, validating their use as proxies for safety assessment. This technique avoids the limitation of conventional strategies that depend entirely on sparse crash data. The research additionally highlighted the significance of incorporating driver behavior, including but not limited to speed and braking behavior, into the safety evaluation. This enriched data allowed for more comprehensive models which can capture the complex interactions among driver behavior, traffic conditions, and safety outcomes. The findings show the value of using the wealth of information through connected vehicle technologies to enhance road safety evaluation. Statistical techniques, such as Negative Binomial regression and Extreme Value Theory (EVT) modeling, have been implemented to analyze the surrogate safety measures and predict the chance of extreme safety events. Implementation of these advanced methodologies offers a more rigorous and complicated technique to safety assessment. Furthermore, the incorporation of a safety index into a route-selection algorithm highlighted the potential for connected vehicle data to inform and optimize transportation decision-making. By guiding drivers toward safer routes, this technique shows how safety considerations could be systematically integrated into navigation systems. Despite the above contributions, the dissertation also recognized several obstacles that need to be addressed in future studies. These include but not limited to data quality and completeness issues, data processing algorithm issues, limited crash types in the safety measures, and the challenge of low connected vehicle penetration rates. Recommendations for future work include improving data collection methods, enhancing the data processing algorithms, expanding the scope of the safety measures and analysis, and developing effective strategies to address the low connected vehicle penetration rates. By addressing those obstacles and pursuing those future research directions, the field of surrogate safety measures in road safety studies, especially with more comprehensive connected vehicle data, can keep evolving, supplying increasingly robust and actionable insights to improve road safety.
Recommended Citation
He, Zhaoxiang, "COMPREHENSIVE ANALYSIS, EVALUATION, AND APPLICATION OF SAFETY SURROGATE MEASURES" (2024). Theses and Dissertations. 3475.
https://dc.uwm.edu/etd/3475