Abstract :
[en] Traffic congestion is a persistent issue that most of the people living in a city have to face every day. Traffic density is constantly increasing and, in many metropolitan areas, the road network has reached its limits and cannot easily be extended to meet the growing traffic demand. Intelligent Transportation System (ITS) is a world wide trend in traffic monitoring that uses technology and infrastructure improvements in advanced communication and sensors to tackle transportation issues such as mobility efficiency, safety, and traffic congestion. The purpose of ITS is to take advantage of all available technologies to improve every aspect of mobility and traffic. Our focus in this thesis is to use these advancements in technology and infrastructure to mitigate traffic congestion. We discuss the state of the art in traffic flow optimization methods, their limitations, and the benefits of a new point of view. The traffic monitoring mechanism that we propose uses vehicular telecommunication to gather the traffic information that is fundamental to the creation of a consistent overview of the traffic situation, to provision real-time information to drivers, and to optimizing their routes.
In order to study the impact of dynamic rerouting on the traffic congestion experienced in the urban environment, we need a reliable representation of the traffic situation. In this thesis, traffic flow theory, together with mobility models and propagation models, are the basis to providing a simulation environment capable of providing a realistic and interactive urban mobility, which is used to test and validate our solution for mitigating traffic congestion. The topology of the urban environment plays a fundamental role in traffic optimization, not only in terms of mobility patterns, but also in the connectivity and infrastructure available. Given the complexity of the problem, we start by defining the main parameters we want to optimize, and the user interaction required, in order to achieve the goal. We aim to optimize the travel time from origin to destination with a selfish approach, focusing on each driver. We then evaluated constraints and added values of the proposed optimization, providing a preliminary study on its impact on a simple scenario. Our evaluation is made in a best-case scenario using complete information, then in a more realistic scenario with partial information on the global traffic situation, where connectivity and coverage play a major role. The lack of a general-purpose, freely-available, realistic and dependable scenario for Vehicular Ad Hoc Networks (VANETs) creates many problems in the research community in providing and comparing realistic results. To address these issues, we implemented a synthetic traffic scenario, based on a real city, to evaluate dynamic routing in a realistic urban environment. The Luxembourg SUMO Traffic (LuST) Scenario is based on the mobility derived from the City of Luxembourg. The scenario is built for the Simulator of Urban MObiltiy (SUMO) and it is compatible with Vehicles in Network Simulation (VEINS) and Objective Modular Network Testbed in C++ (OMNet++), allowing it to be used in VANET simulations.
In this thesis we present a selfish traffic optimization approach based on dynamic rerouting, able to mitigate the impact of traffic congestion in urban environments on a global scale. The general-purpose traffic scenario built to validate our results is already being used by the research community, and is freely-available under the MIT licence, and is hosted on GitHub.
