Study of the effects induced by the introduction of automated vehicles on vehicular traffic flow in urban areas

In recent years, the interest of car manufacturers in the development of self-driving vehicles has grown, and automated vehicles are expected to circulate in increasing numbers in the next years. These predictions lead to the need to analyse how automated vehicles will interface with conventional vehicles and existing infrastructure. Given the small number of automated vehicles in circulation, the available literature proposes various tools for modelling them within microsimulation software and shows results in terms of system efficiency referring to highway scenarios. Few publications focus on urban scenarios and no publication concentrates on the exclusive analysis of a lane change maneuver. This thesis aims to close this gap in the literature by analyzing how the lane change behaviour of automated vehicles impacts efficiency and safety within the vehicular flow in an urban network. To achieve this goal, initially the urban network of the city of Hannover, Germany, was modeled and two different demand profiles were loaded. The first one refers to the profile of a typical working day, while the second profile represents a fictitious demand, obtained by increasing and subsequently decreasing the number of vehicles in circulation. An analysis of the available literature was then carried out to find parameters with which to model both automated and conventional vehicles. Eleven different scenarios are created, with the reference Scenario 0 consisting of the circulation of conventional vehicles only, in which the number of autonomous vehicles in circulation is increased by 10% for each subsequent scenario. The final scenario consists of only autonomous vehicles in circulation within the network. The results and final considerations are presented in terms of efficiency, i.e. through the analysis of travel times, average travel speeds and through the analysis of fundamental network diagrams, and in terms of safety with reference to Scenario 0. The results show a moderate but statistically significant worsening in both efficiency and safety. Results presented in this thesis can form the basis for further research aimed at studying the actual lane change behaviour of automated vehicles by developing ad-hoc models, or aimed at improving the efficiency and safety of lane change maneuvers of automated vehicles.