Abstract
dc:descriptionThe current doctoral dissertation examines the performance evaluation and optimization of Vehicle-to-Everything (V2X) communications. The main objective is to identify the key factors that affect the performance of V2X networks as well as the Quality of Service (QoS) offered to vehicular users. For this purpose, it introduces innovative strategies and closed-form expressions that can lead to system’s service provisioning improvements and more accurate evaluation of V2X network’s performance. With a view to providing a complete and definite understanding of the factors shaping V2X performance, this dissertation explores the following three distinct research areas. The first research area investigates the impact of vehicular mobility on the service quality in vehicular WiFi networks serving passenger requests. Specifically, the adoption of Bandwidth Reservation (BR) policy is proposed to prioritize the service of handover calls. Two types of BR policy are considered: i) the strict BR policy, where the reserved resources are fixed and ii) the probabilistic BR policy, where the number of reserved resources is determined dynamically by a probabilistic metric. To further improve service provision to handover calls, a waiting queue is exclusively dedicated to handover calls. For the system performance evaluation along with the employed policies, three novel iterative algorithms are proposed and the solutions are compared against the conventional Complete Sharing (CS) policy, in which no priority exists between the users. The second research area focuses on the performance assessment of V2X communications in scenarios where direct connectivity between vehicles and the cellular network, i.e. Vehicle-to-Network (V2N) is not available. Classical queueing theory models, such as M/M/C, M/Ekp/C και M(N)/M/C are utilized to evaluate how the service-time distribution and the vehicle population affect the system’s performance, while accounting for network impairments such as channel failures. The impact of direct and indirect V2N communications on in-vehicle hosted applications is assessed and service improvement strategies are introduced and validated through closed-form analytical expressions. The third research area integrates Artificial Intelligence (AI) methods in V2X systems. This contribution is two-fold: First, novel hybrid performance evaluation approaches are developed that combine classical methods such as closed-form expressions and/or simulation with AI-driven techniques to create an efficient and accurate performance assessment framework. Second, the deployment of distributed AI is examined, specifically Federated Learning (FL), which is poised to play a pivotal role in V2X systems due to its advantages in preserving user data privacy. Given that the successful training of FLmodels heavily relies on the collaboration among multiple vehicles, an overview of network metrics affecting FL is presented, alongside with a closed-formanalysis to determine whethernetwork conditions can ensure the seamless training of the FL algorithm.
Degree
thesis:*- Grantor dc:publisher
- University of Peloponesse
- Year dc:date
- 2025
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Keramidi, Eirini-Lydia
- Κεραμίδη, Ειρήνη-Λυδία
Subjects
dc:subject × 22- Επικοινωνίες V2X
- Ασύρματα δίκτυα οχημάτων
- Αξιολόγηση απόδοσης
- Θεωρία ουρών αναμονής
- Υβριδικές μέθοδοι εκτίμησης απόδοσης
- Διαχείριση πόρων δικτύου
- Τεχνητή νοημοσύνη
- Ομοσπονδιακή μάθηση
- V2X communications
- Vehicular networks
- Performance evaluation
- Queueing theory
- Hybrid performance estimation methods
- Network resource management
- Artificial intelligence
- Federated 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.*- Handle dc:identifier
- http://hdl.handle.net/10442/hedi/61128