The evolution of Intelligent Transportation Systems (ITS) represents a significant shift towards enhancing transportation management and infrastructure through advanced communication technologies. ITS aims to improve road safety, traffic flow, and driving comfort, but faces challenges with latency, data processing, and service continuity due to the demanding nature of its applications. Edge Computing emerges as a key solution, enabling localized data processing close to the source, thus reducing latency and improving the efficiency of vehicular applications in ITS.
This thesis explores the integration of Edge Computing into vehicular networks, addressing the complexities involved, such as edge server placement, efficient data offloading, service migration protocols, and ensuring privacy and security. The research is three-pronged: Firstly, it introduces the Edge-based Safety Intersection Assistance System (ESIAS) to enhance intersection safety by distributing timely warnings to drivers, reducing accident risks. Secondly, it investigates the optimal placement of Edge servers within vehicular networks through integer linear programming, considering factors like latency, cost, and server capacity to minimize deployment costs and balance computational loads while keeping latency in check. Thirdly, the thesis addresses service migration in mobile edge computing (MEC)-enabled vehicular networks. As vehicles move, quality of service (QoS) maintenance requires strategic service migrations, which are challenging due to timing and location. To overcome these challenges, the thesis formulates service migration as a Markov Decision Process (MDP) and employs deep reinforcement learning, specifically Deep Q-Networks (DQN), to develop optimal migration strategies. These strategies are designed to meet each service's specific needs, ensuring uninterrupted service continuity even in high-mobility environments, striking an essential balance between reducing latency and migration costs. Overall, the thesis contributes to the advancement of ITS by integrating Edge Computing, enhancing the performance and reliability of vehicular networks.

Amphi LaBRI