Doctoral Dissertations

Orcid ID

https://orcid.org/0000-0001-5026-8644

Date of Award

8-2025

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Computer Engineering

Major Professor

Himanshu Thapliyal

Committee Members

Himanshu Thapliyal, Qing Cao, Jinyuan Sun, Annarita Giani

Abstract

The rapidly growing field of quantum computers has brought forth a revolution in the world of computing for the past decade. However, with this comes immense potential for change that breaks down the very foundation of what was previously considered computationally infeasible. In a quantum-enabled world, the fundamentals of cybersecurity change immensely, from now-unsafe cryptography to quantum-secured data. This work aims to investigate both by harnessing the power of quantum computing in one of the most high-risk, vulnerable domains in modern society: cyber-physical systems. To achieve this, we take a look at the taxonomy and structure of a typical CPS ecosystem to determine the potential weaknesses and gains from quantum computing.

First, we analyze the physical-level security from the perspective of securing Controller Area Network-based (CAN) electronic control units, essential components of modern vehicles and machinery, to develop an efficient, quantum-safe framework for trusted device authentication. Our authentication scheme featuring post-quantum cryptography (PQC) uses 4 times fewer messages than leading vehicular security frameworks. Next, we extended this work to inter-device CPS communication in the context of additive manufacturing networks using the Controller Area Network (CAN) bus. Our proposed design provides quantum-safe, root-of-trust tree networks to 3D printing farm networks, allowing plug-and-play connections without fear of trojan-horse or man-in-the-middle attacks. Using 12 times fewer messages for authentication compared to other PQC systems, our framework defends against most forms of man-in-the-middle attacks and effectively mitigates most denial-of-service attacks. Furthermore, we created a supervisory-level anomaly detection model utilizing quantum fidelity kernel support vector machines, providing up to a 13% boost in classification performance of real-world industrial control system data and an 11% in 3D printer anomaly data. Utilizing the robust geometric distance metric, we found a clear, quantifiable advantage to quantum machine learning for securing CPS data and feedback systems. Finally, we discuss the future directions of research that could build upon this multi-level framework. Our work successfully bridges the gap between quantum computing in cyber-physical systems, both as a potential threat and strength to the future of critical infrastructure, supply chain trust, and manufacturing safety.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS