Date of Award

12-2024

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical and Civil Engineering

First Advisor

Kim-Doang Nguyen

Second Advisor

Thomas C. Eskridge

Third Advisor

Hector M. Gutierrez

Fourth Advisor

Seong Hyeon Hong

Abstract

This dissertation advances the development of robust learning-based approaches across two complementary domains: engineering education and autonomous systems. Through four studies, this research addresses critical challenges in preparing data-proficient engineers and developing reliable autonomous systems that can operate under uncertainty and incomplete information. The engineering education study examines how mechanical and aerospace engineering undergraduates conceptualize and develop data proficiency skills essential for modern engineering practice. Through interviews with 27 students, the research employs the How People Learn framework to analyze student perspectives on information literacy, data interpretation, and computational thinking. The findings inform pedagogical strategies for developing data proficiency in engineering education, ensuring graduates are better equipped for an increasingly data-driven profession. The subsequent robotics studies investigate three distinct aspects of autonomous system robustness. The first study evaluates the resilience of deep Q-learning algorithms in autonomous driving scenarios under cybersecurity threats. Using the CARLA simulator, the research reveals that deep Q-network agents demonstrate inherent resilience to denial-of-service attacks by adopting cautious behaviors, while showing vulnerability to deception attacks targeting semantic image inputs. The second study addresses robust UAV landing on moving platforms under partial observability, introducing a novel Robust Policy Optimization algorithm integrated with Long Short-Term Memory networks (RPO-LSTM). This approach significantly outperforms traditional methods when handling sensor failures and noisy data during landing maneuvers. The final study presents LEVIOSA, an innovative framework that leverages multimodal Large Language Models to generate UAV swarm trajectories from natural language commands. The framework incorporates multi-critic consensus mechanisms and hierarchical prompting to ensure accurate and synchronized multi-UAV coordination. Collectively, these studies make significant contributions to both theoretical understanding and practical implementation of robust learning-based systems. The engineering education research provides evidence-based insights for developing effective data proficiency curricula, while the robotics studies advance the field through novel algorithms and frameworks that enhance autonomous system reliability under challenging conditions. Together, these contributions support the development of more resilient autonomous systems and better-prepared engineers, addressing crucial needs in both education and technology sectors. The dissertation demonstrates the value of pursuing robustness across multiple domains, offering solutions that bridge the gap between engineering education and advanced autonomous systems development.

Download

Share

COinS