The social impact of urban traffic can be assessed in many ways. Quantitatively speaking, there are estimates that the loss of time in traffic corresponds to billions of dollars per year. However, there are other dimensions, given that better mobility is an important force for the alleviation of poverty. Both facts have been known by traffic authorities; in fact, they actively use the most effective techniques from traffic engineering. However, the use of Internet has evolved little in this area, especially in what concerns communication within the traffic infrastructure. Moreover, the current trend in this regard is centered around private stakeholders, with little synergy with the official traffic authorities in charge. This happens in spite of the fact that the presence of mobile devices and Internet in this scenario is ubiquitous. The technology is already mature for the so-called mobility Internet, thus cars and other traffic participants are able to exchange information. Traffic authorities could then collect, perform analytics and use such information in various ways and time horizons, as e.g., for short-term prediction and for policy-making. Extracting this potential from all the elements in the traffic system, as well as from diverse data sources requires new paradigms of data gathering, broadcast, simulation and optimization, including those related to distributed artificial intelligence and machine learning. This project thus proposes the development of methods and tools for investigating the use of Internet in various traffic-related contexts, such as route choice by connected vehicles, and intelligent traffic signal controllers.
The goal of this project is twofold. First, it aims to investigate the effects of interconnecting the various actors (driver-vehicle units, the traffic infrastructure), in regard to the task of reducing traffic congestion and improving life quality for societies. Secondly, I aim at developing new AI-based approaches for optimization in traffic scenarios.