Spatial big data analysis and artificial intelligence applications for transportation safety and network systems optimization
No Thumbnail Available
Authors
Meeting name
Sponsors
Date
Journal Title
Format
Thesis
Subject
Abstract
In an era of rapid urbanization and increasing mobility, transportation safety and efficient network systems are paramount for sustainable urban development. This dissertation delves into the intersection of spatial big data analysis and artificial intelligence (AI) to address critical challenges in transportation safety and network optimization. Leveraging large-scale spatial data sources, including GPS traces, traffic camera feeds, and geographical information systems (GIS) data, this research employs advanced machine learning algorithms and AI techniques to enhance transportation safety, improve traffic management, and design adoptable frameworks geared towards optimizing network systems. The first part of this dissertation dives into the realm of surrogate cash measures derived from connected vehicle data. Leveraging the rich information obtained from connected vehicles, this research examines the feasibility and effectiveness of using surrogate measures, such as traffic flow characteristics and driver behavior patterns, as proxies for assessing transportation safety. Advanced spatial analysis models are employed to extract and analyze these surrogate measures, providing insights into accident risk prediction and proactive safety interventions. The second part of the dissertation shifts its focus to the acceleration of the connected vehicle big data pipeline. Recognizing the sheer volume and velocity of data generated by connected vehicles, this research investigates strategies to streamline data collection, processing, and analysis. Cutting-edge techniques in data engineering and distributed computing are employed to create an efficient and scalable pipeline, ensuring timely access to real-time data for traffic management and decision-making. The third part focuses on predictive models for accident analysis and risk assessment. By analyzing historical accident data, spatial patterns, proxy driving behavior metrics, and traffic parameters, machine learning models are designed to predict accident hotspots and identify contributing factors, which is capable of detecting crash outcomes from traffic state information extracted from CV data at 5 min intervals, with accuracies ranging from 88-95 percent. These predictive models offer the potential to proactively mitigate safety risks and allocate resources effectively. In the final part, this dissertation delves into trajectory data analysis extracted from traffic video feeds. With the proliferation of traffic cameras, video data has become a valuable resource for understanding vehicle movements and traffic dynamics. Advanced computer vision algorithms and AI techniques are harnessed to extract and analyze trajectory data, enabling improved traffic management, by assessing location specific risk factors based on the characteristics of the traffic as it interacts with the infrastructure and with each other. Throughout this dissertation, a multidisciplinary approach is adopted, drawing from the fields of computer science, data science, transportation engineering, and urban planning. By harnessing the power of spatial big data and AI technologies, this research contributes to the advancement of transportation safety, network efficiency, and sustainable urban mobility. In conclusion, the findings and methodologies presented in this dissertation offer valuable insights into the transformative potential of spatial big data analysis and artificial intelligence applications for enhancing transportation safety and network systems. This work has far-reaching implications for urban planners, transportation authorities, and policymakers striving to create safer, more efficient, and sustainable transportation systems in the modern urban landscape.
Table of Contents
DOI
PubMed ID
Degree
Ph. D.