Effects of Vehicles with Different Degrees of Automation on Traffic Flow in Urban Areas

Abstract

In addition to the electrification of the drive train, the development and use of autonomous and highly automated vehicles is another important contribution to the future development of individual transport. The automation of vehicle guidance is seen as a means to avoid traffic accidents, to gain leisure time in automated driving phases, but especially to improve the traffic flow and to avoid traffic jams as far as possible. While there are many obstacles to the popularization of fully autonomous vehicles in most countries, vehicles with driver assistance systems up to level 2 are already very common in different regions. A better understanding of the effects of vehicles with higher levels of automation on traffic flow can provide positive impulses for political strategies, the expansion of urban infrastructure, and the development of individual vehicles. In this thesis, the most important driving differences between fully automated vehicles, semi-automated vehicles, and non-automated vehicles are investigated and corresponding vehicle driving models are developed. To compare the effects of these vehicle driving mod- els in real traffic, four microscopic traffic simulation scenarios are created for two cities in two countries with different infrastructure. When comparing the simulation results for traf- fic with vehicles of different degrees of automation, simulation results of vehicles with a high degree of automation show a better influence on traffic in terms of traffic density, average speed, and travel time. Vehicles with high automation degree can improve the traf- fic situation from 3.7% to 57.4% in different aspects. Higher automation degree can bring greater impact on traffic. Based on the current development of automated vehicles in terms of technology, politics, price, and public opinion, this work estimates the penetration rate of auto-mated vehicles in the near and far future. The future traffic status is simulated based on the different penetration rates of automated vehicles. Even if only a part of the vehicles in the traffic flow is automated, this can have a positive effect on the traffic flow in many ways. The traffic scenarios of the German city of Duisburg and the Chinese city of Wuhan created in this thesis are based on real road networks and traffic data. The scenarios can realistically reflect real traffic conditions and can also be used for further work in traffic research. The vehicle guidance model built in this thesis is based on driving experiments in a driving simulator, but it can also be combined with different vehicle models to simulate the effect of combining different drivers and vehicle types.