Electronic Theses and Dissertations

Date

2024

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Computer Science

Committee Chair

Myounggyu Won

Committee Member

Kan Yang

Committee Member

Sabyasachee Mishra

Committee Member

Weizi Li

Abstract

Traffic congestion is a serious problem in the USA that affects safety, economy, environments, and human lives. Autonomous vehicles (AVs) equipped with vehicle-to-everything (V2X) communication technology is emerging as a viable solution to mitigate traffic congestion. In this dissertation, we propose an advanced traffic control system for autonomous vehicles, utilizing machine learning techniques, to alleviate traffic congestion, and enhance traffic efficiency and safety. The proposed system consists of two key components: an intelligent adaptive cruise control system (ACC) and a cooperative lane-change system. To address the limitations of existing static model-based approaches, we introduce a novel AI-based ACC system that dynamically adjusts the ACC settings based on real-time traffic conditions. By adapting to changing situations, this system significantly improves traffic efficiency. However, we recognize that current intelligent ACC systems primarily focus on traffic flow enhancement, disregarding the influence of adaptive inter-vehicle gap adjustment on driving safety and comfort. To bridge this gap, we develop a Safety-Aware Intelligent ACC system, which effectively assesses driving safety by dynamically updating safety model parameters according to varying traffic conditions. This innovative approach ensures that driving safety and comfort are prioritized alongside traffic efficiency. Furthermore, we present a novel multi-agent reinforcement learning (MARL)-based intelligent lane-change system for autonomous vehicles. This system optimizes both local and global performance by incorporating a road-side unit (RSU) responsible for managing a specific road segment, as well as vehicle-to-everything (V2X) capabilities for the agents. This density-aware cooperative multi-agent framework enables efficient and safe lane changes, considering the overall traffic conditions and maximizing the benefits for all vehicles involved. Finally, we present a use case scenario of our proposed next-generation traffic control system by designing an intelligent adaptive motion control system for electric vehicles (EVs) which facilitates an EV to control its motion to align with the position where the electromagnetic strength is expected to be maximal to receive maximum charging efficiency. By combining the AI-based ACC system and the MARL-based intelligent lane-change system, our next-generation traffic control system for autonomous vehicles aims to revolutionize traffic management, offering improved efficiency and safety for autonomous vehicles on the roads of the future.

Comments

Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to ProQuest.

Notes

Embargoed until 07-10-2025

Download

Available for download on Thursday, July 10, 2025

Share

COinS