June 2017 – December 2019
In the SAFARI research project, the state of Berlin and its partners from administration, research and industry are testing the exchange and updating of digital maps as one of the basic requirements for automated and connected driving. Within 31 months, the partners involved have jointly developed and tested technologies of highly automated and connected driving in an urban scenario. The project was funded with €4.3 million by the Federal Ministry of Transport and Digital Infrastructure BMVI within the framework of the funding guideline “Automated and Connected Driving on Digital Test Fields in Germany”.
In addition to the Senate Department for the Environment, Transport and Climate Protection (SenUVK), the Senate Department for Economic Affairs and Energy (SenWEB) and the Reinickendorf district office participated in the project on behalf of the state of Berlin. Partners in the project were the Free University of Berlin, the Technical University of Berlin with its associated institute DCAITI and the research institute Fraunhofer FOKUS, the companies IAV and HELLA Aglaia, and Deutsche Telekom as an associated partner.
As part of SAFARI, various vehicles competed to test the reliable and safe use of sensing technologies and derived data uses:
The initial focus was on the more precise so-called self-localization of the AVF. Landmarks can be extracted from the open databases of the state of Berlin with precisely measured objects of the traffic infrastructure, the light poles and also the street trees. Orientation on these was tested with research vehicles to find out what minimum requirements for self-localization and thus orientation in urban traffic must be placed on autonomous series-produced vehicles in urban traffic.
The goal of the state of Berlin must be to ensure the greatest possible safety of autonomous driving. The demanding task of optically detecting light signals and adapting driving behavior to them is supported by the additional transmission of signal information and information about the intersection topology to the vehicles. With Vehicle2Infrastructure technology, direct communication from the traffic light systems to the vehicles is implemented and the effects of traffic stabilization and improved efficiency of traffic performance are also tackled. The integration of the new technology into the existing traffic technology and the proof of its reliability were the focus of the testing.
Automated vehicles should also be able to park autonomously in the target area. In SAFARI, tests were carried out to see how good the vehicles are at recognizing free unmarked parking spaces at the side of the road, even when driving past. In comparison with the digital maps, an accurate picture of free parking spaces can be generated and constantly kept up to date, effectively reducing parking search traffic.
Of course, automated vehicles are trained to detect obstacles. Transmitting the location of current road works to the vehicles helps them to detect these obstacles earlier and improves driving comfort. However, before these vehicles drive nationwide, a significant improvement in traffic management can be achieved even with a low level of automated vehicle equipment.
If objects of the road infrastructure, such as traffic signs or markings, are missing, automated vehicles should pass on this “error”. For the administration, every automatically generated indication of such “errors” is valuable. This includes changes and damage to these objects, as well as to the road surface. The testing of the so-called change detection across the collected indications of several vehicles and the suitable provision to the administration, e.g. for road maintenance, help to quickly identify and repair missing or defective infrastructure elements.
n addition, what if the map used as a reference was not up to date? he State of Berlin has long been committed to the digitization of the administration. For the Senate Department for the Environment, Transport and Climate Protection, digital maps are already an indispensable basis for work. With the step towards automated detection of missing objects or changes and the development of a continuous automated updating of the map basis, digitalization is being consistently advanced. Of course, the automated vehicles also benefit when “errors” in the map are quickly detected and eliminated.
inally yet importantly, second row parking also challenges automated vehicles. Knowing this restriction of the available road space in a timely manner allows the vehicles to plan a safe and comfortable journey through greater foresight. What for the AVF is merely a dynamic object in the road space and only valid for a short time, on the other hand, serves the administration to identify performance limitations and to ensure the increase of road safety.
The high-precision map of the state of Berlin created in previous projects is used as a reference map in the project. In addition, it was tested how it can be extended with dynamic georeferenced information as a Local Dynamic Map (LDM). The LDM maps the dynamic traffic situation and the actual state of the surroundings, including information on road works or hazardous situations. The linking of all information allows the highly accurate perception of the surroundings (localization, free space and object recognition) and enables intelligent behavior prediction. In this way, complex, dynamic and often unstructured traffic conditions can become controllable for the vehicles even beyond the range of their own sensors or mapped surroundings.
Finally, the secure AVF relies on high availability of mobile communication between vehicles and with the infrastructure. New suitable mobile radio standards (5G) and a new distributed communication infrastructure (Small Cells) are currently being developed and tested. In SAFARI, different variants were investigated in search of the technically and economically most favorable solution. In all these tasks, the standards already developed in the field of communication and data exchange were used. The state of Berlin is thus integrating itself into national and European developments in preparing the municipal infrastructure for the AVF.
This is another reason why the state of Berlin is promoting the use of the conditions created in the municipal digital urban traffic test field in the Reinickendorf district for other automated and connected driving projects in the state of Berlin. In doing so, the necessary adaptations and expansions of the infrastructure resulting from these projects can be identified and implemented.