The main objective of this deliverable is to study various models in land transport domain and various ways of modelling. Chapter 1 "Existing Models For Land Transport Using Intelligent Vehicles" describes an existing approach in the domain that presents some similarities with our project. The objective of this work is to identify which concepts are relevant for TACOS, which ones are out of the scope of the project and which ones are lacking in relation to TACOS requirements. Chapter 2 "A Model of the Platooning Problem in Classical B" tackles the particular problem linked to the construction of a platoon of autonomous vehicles. Up to now, only the property of non-collision has been considered. The approach consists in using a mathematical model of the behavior of the vehicles and specifying how the platoon moves while satisfying the given property. Chapter 3 "Localization Component: Relevant Properties and First Algorithms" gives a first attempt to specifying an important component for land transport: the localization component. The approach consists in defining relevant properties of this component and to study whether they can be considered or not by the localization algorithm formalized with the B language. The two following chapters do not specifically deal with models in land transport domain but presents methodological aspects allowing the definition of a design method adapted to the domain and its support tool. Chapter 4 "Extending KAOS to support variability" starts from the observation that there is not a single Cycab system but several systems that share common properties and own specific properties. The idea is then to define a concept of variability in the design method. As KAOS was chosen in the project, the chapter proposes to extend it to take into account this concept. Chapter 5 "Ecriture d'un métamodèle pour KAOS. Intérêts et objectifs de ce métamodèle" aims at evaluating a particular approach to define a tool allowing, among others, the method described in chapter 4 to be implemented. This approach is based on the concept of metamodels used in MDE (Model Driven Engineering).