In the past few years the progress in the development of autonomous driving vehicles has increased tremendously. There are still technical, infrastructural and regulative obstacles which have to be overcome, but fully autonomous driving vehicles will be reality in our cities in the coming years or at least in the coming decades.
Most likely it will be cheaper and more convenient to utilize autonomous driving vehicle fleets than maintaining privately owned cars, so a big shift to carsharing will go on.
While the economic, ecologic and social impact of widely used autonomous driving fleets are tremendous, there are hardly any data and studies for the effects on cities and municipalities available.
To meet this demand, the University of Applied Sciences in Esslingen, Germany took part on the research project “KI4ROBOFLEET“.
In this project a SUMO based simulation environment for autonomous vehicle fleets was implemented.
With the “Scenario Builder“ users can easily create simulation scenarios for autonomous driving fleets, including point-of-interest based scenarios like leisure (going to the theaters or sporting events) or work. The simulation takes place in the city map of Mannheim with real sensor based traffic data.
An example is a "Morning Scenario" where people are picked up at home by autonomous vehicles and are chauffeured to offices, to schools, to stations, etc.
The simulations are performed iteratively with different fleet sizes and routing strategies.
After the simulation iterations are completed the results can be evaluated by economic and ecologic parameters. Also the passengers waiting time can be determined and so a well-balanced fleet size can be found for a certain amount of customer requests.
This tool can support city planners and carsharing companies to evaluate the chances and risks of running autonomous driving fleets in their local environment with their characteristic infrastructure.

свернуть