Access Type

Open Access Dissertation

Date of Award

January 2025

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Computer Science

First Advisor

Weisong Shi

Second Advisor

Zheng Dong

Abstract

Road traffic accidents are a significant global concern, claiming approximately 1.3 million lives annually and causing non-fatal injuries to 20–50 million people, many of which result in long-term disabilities. They are the leading cause of death for individuals aged 5–29 and impose a substantial economic burden, costing most countries 3% of their Gross Domestic Product (GDP). In addition to car accidents, other types of road incidents, such as collisions involving tall vehicles and overpasses, also pose significant risks. These accidents result in numerous fatalities and cause millions of dollars in damages annually. Environmental conditions like wet or icy roads and poor visibility further increase the likelihood of accidents. Crashes are particularly frequent during rush hours and late-night hours, with intersections being especially hazardous, accounting for over 50% of urban road accidents. Weather also plays a significant role, with rain increasing crash rates by up to 34% and snow or ice exacerbating the risks. These trends underscore the urgent need for targeted interventions to improve road safety worldwide. Our research focuses on innovative approaches to minimize accidents and save lives. First, we propose SafeOverPass, a low-clearance warning system that uses edge computing and machine learning to provide critical alerts to drivers of tall vehicles. Second, we developed a method for determining a vehicle’s position using real-time data from an On-Board Diagnostic II (OBD-II) Vehicle Interface, addressing challenges posed by GPS signal loss. Third, we implemented a data-driven approach that integrates historical intersection accident data with real-time traffic information to reduce the likelihood of intersection accidents along a vehicle’s route. Lastly, recognizing the importance of rapid emergency response, we designed a mathematical model leveraging historical and real-time data to optimally position ambulances at and near intersections, minimizing response times and improving survival outcomes.

Download

Share

COinS