Enhancements of Autonomous Vehicles Environment based on smart communication systems

Abstract

The first part of this thesis is focused on the vehicular networks and on the different protocols that are used in vehicular environment to exchange information and data. A description of the most used protocols of Vehicular Ad-hoc NETwork (VANET) is presented along with their positive and negative aspects, their issues and some improvements. The capability of exchange data and information is one of the most important features to improve everything making objects smart. Obviously this is valid also for vehicles that can be improved using connectivity and networks. New paradigms and architecture are growing in interest thanks also to new technologies such as lightway protocols or 5G giving birth to new Internet of Vehicles (IoV). Autonomous vehicles are a special kind of vehicles able to move in an urban environment without any human interaction thanks to several sensors equipped onboard. These sensors can detect obstacles and possible dangerous situations on the road during a journey. To ensure a high reliability it is necessary a cooperation between all equipped sensors like cameras, Lidar, Radar and GPS. Image and video elaboration is made through computer vision algorithms that are able to identify and track objects moving inside sensors line of sight while lidar and radar create a threedimensional model of the surrounding environment. These gathered information have only a local scope. In order to solve this problem a “collaborative” approach could be used to establish a communication network among these vehicles. In this way every autonomous vehicle could be able to handle warning situations and traffic jam in a better way exploiting information exchange with others. It is important to build a robust network infrastructure in order to achieve a reliable message exchange. Vehicles , equipped with on board units, are starting to be considered part of the IoT domain, where instead of classical smart objects we consider smart vehicles. These smart vehicles can exchange messages among themselves to improve their useful environment knowledge. New routing protocols and optimization algorithms are necessary in order to make the network infrastructure more flexible and able to adapt to changes. Furthermore vehicles move at a very high speed in comparison to other common devices and the link low stability makes the network high dynamic and unstable. To resolve these issues it will be necessary to investigate actual VANET protocols and improve them to better manage its high dynamism and link quality using also technologies that belong to Internet of Things domain. Integrating these heterogeneous technologies together it will be possible the creation of a middleware between the IoT layer and On Board Units. These solutions, based on evolutive algorithms and adaptive protocols have to be verified through simulations and mathematical models. To do so, a deep study on mobility models and simulation tools has been necessary. The future of mobility is, taking into consideration the recent trends, electric. Fuel propelled vehicles will some day disappear from our city to leave plate to this new kind of vehicles with electric engine and powered by batteries. Electrical mobility will be a very important part of the future and also autonomous vehicles will be electric. The consumption of on-board vehicle sensors must be taken into account and must be modeled by simulation tools that are used to describe future mobility scenarios. In the end a possible application of all studied research fields is proposed, summarized inside the proposition of a platooning application for autonomous vehicles. A platoon consists of a set of vehicles that moves at the same speed along the road with a low inter-vehicle gap. In this the members of the platoon will benefit of several advantages that will be treated in the body of this thesis.