Harokopio University
AI‐enabled management of next‐generation autonomous transport systems
Abstract
dc:descriptionIn recent years, the modern world has witnessed a significant increase in the use of automated systems in various areas, motivated by the desire for greater efficiency and sustainability. This trend is fueled by technological advancements and the support of Information and Communication Technologies (ICT), which have greatly enhanced the complexity and capabilities of these systems, particularly in the transportation sector. The design and implementation of transportation solutions have significantly accelerated the development of intelligent and autonomous transportation systems and the provision of many innovative services. Under this scope, the drive to replace human operators with autonomous agents in complex and challenging tasks, with the ultimate goal of improving efficiency, safety, and reliability, is fundamentally transforming the landscape of mobility across a wide range of applications. At the core of this evolution lies the integration of Artificial Intelligence (AI) and Machine Learning (ML) methods, which are propelling vehicles toward unparalleled levels of autonomy and intelligence by emulating human‐like reasoning processes. Leveraging advanced AI and ML with a particular emphasis on Deep Reinforcement Learning (DRL) techniques can facilitate faster decision making by automating the real‐time sensing, detection, response, and control processes in intelligent and autonomous vehicles. This progress is fueled by the pursuit of greater efficiency, enhanced safety, and improved reliability for a plethora of scenarios and applications across diverse domains, including aerial, underwater, and ground environments. The deployment of AI in real‐time sensing, decision making, planning, and managing provides a transformative solution for these autonomous systems, enabling them to maximizing operational efficiency and minimizing resource usage in various operational conditions, that cannot be effectively handled with traditional methods. Commencing with an overview of the current industry landscape and with respect to sustainability, adaptability, and human‐centric design, this Ph.D. dissertation delves into the fundamental role of AI and ML methods and algorithms in shaping the decision making capabilities of autonomous and intelligent vehicles, aiming to significantly bolster their operational efficiency. This study adopted a systematic benchmarking approach to assess the capabilities of intelligent and autonomous systems by comparing the performance of various algorithms in controlled as well as dynamic real‐world scenarios. By analyzing traditional path‐planning algorithms such as the A* and Dijkstra alongside adaptive Reinforcement Learning (RL)‐based approaches, the research highlighted the advantages of RL models in enhancing efficiency and reliability. The findings indicate that RL methods can significantly improve the ability of autonomous systems to adapt and learn from their environments, thereby optimizing their overall performance in dynamic situations. In addition, the dissertation emphasizes the societal impact and regulatory considerations of autonomous operations, underscoring the critical need for ongoing development to foster the acceptance, adoption, and application of autonomous vehicles and related technologies in various environments. By addressing these challenges, it establishes a foundational framework for the safe, effective, and seamless integration of autonomous systems into both public and commercial domains, paving the way for their widespread deployment and long‐term adoption. As automation advancements in the vehicle industry continue to progress rapidly, the journey towards fully autonomous and intelligently connected vehicles remains complex and multifaceted, necessitating a convergence of technological innovations alongside and the human perspective. Ultimately, the realization of full autonomy is dependent on the effective collaboration between autonomous systems and human operators. The outcomes of this Ph.D. dissertation tackle essential challenges in autonomous navigation, real‐time decision‐making, and societal integration, presenting innovative solutions aimed at improving system performance and adaptability. By addressing these key areas, the research contributes to the development of more reliable and efficient autonomous systems, facilitating their acceptance and integration into everyday life while ensuring they meet the needs and expectations of society.
Degree
thesis:*- Grantor dc:publisher
- Harokopio University
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Politi, Eleni
- Πολίτη, Ελένη
Subjects
dc:subject × 24- Αυτόνομα συστήματα μεταφορών
- Αυτόνομα οχήματα
- Τεχνητή νοημοσύνη
- Αυτόνομη πλοήγηση
- Αυτόνομα υποβρύχια οχήματα
- Αυτόνομα και μη επανδρωμένα εναέρια οχήματα
- Διασυνδεδεμένα αυτόνομα οχήματα
- Ενιχυτική μάθηση
- Βαθιά ενιχυτική μάθηση
- Autonomous transport systems
- Autonomous vehicles
- Artificial intelligence
- Autonomous navigation
- Autonomous underwater vehicles
- Unmanned aerial vehicles
- Connected and autonomous vehicles
- Reinforcement learning
- Deep reinforcement learning
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Επιστήμες Μηχανικού και Τεχνολογία
- Μηχανική και συστήματα επικοινωνιών, Τηλεπικοινωνίες
- Electrical Engineering, Electronic Engineering, Information Engineering
- Engineering and Technology
- Communication engineering and systems, Telecommunications
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/58448