Development Standards Proposed for Safe Automated Driving:
Results of Joint Project Verification and Validation Methods Presented
Research Focus: Intelligent Transportation Systems and Logistics
The Challenges of Urban Traffic
With pedestrians, cyclists, motorized two-wheelers, badly visible intersections and other vehicles, urban traffic is one of the biggest challenges in automated driving. On top of the number and variety of other road users, the infrastructure with traffic signs and traffic light systems is another obstacle. “In order for the vehicle to handle even extremely rare scenarios in the future, transparent structures and processes are needed that not only enable the safe operation of a system in exceptional situations, but can also demonstrate safe maneuvering,” says Roland Galbas from Bosch, head of the VVM consortium project.
Prerequisite for Traffic Admission: Verifiable Safety
As early as in the design and development of automated driving functions, the safety principle comes first. Accordingly, these safety functions must be proven for the traffic admission of a vehicle and a certified release for road traffic. To provide this evidence, the 21 project partners jointly developed a model consisting of various procedures, methods and tools. In this way, a so-called safety argumentation can be used to prove that the system can be used safely.
For the methodological design of this model, the partners worked together in several subprojects. As the largest scientific partner in the project, the FZI Research Center for Information Technology is responsible for subproject 7 “Test Procedure”. The focus is on the development of comprehensive and consistent testing methods for Level 4 and Level 5 vehicles in urban areas. The developed concepts and methods enable a consistent testing procedure. They range from simulations up to real-world tests for a modular safety argumentation of automated vehicles. In addition, central infrastructures in the form of simulation environments, hardware-in-the-loop test stands and test vehicles were set up and extended at the FZI.
The defined test model, applied industry-wide, would form the basis for proving the safety of automated vehicles. “For the first time, the models developed here make it possible for all car manufacturers to have the same structures at their disposal for the verification and validation of automated driving systems in urban environments. This standardization could then also lead to industry-wide standards that can make road traffic even safer for all road users,” explains Helmut Schittenhelm, project coordinator at Mercedes-Benz.
Technology Pioneer from Germany
The methodological approach from the VVM project is the world’s first standard also taking industrial processes into account. As a result, the German automotive industry once again takes over the role of technology pioneer in automated driving. In 2021, the world’s first regulation for fully automated driving (SAE Level 4) already came into force in Germany by means of a corresponding law. In 2022, a regulation containing the technical details was adopted in order to be able to register and operate the corresponding vehicles on German roads.
At the end of its runtime, building on the results of predecessor projects Pegasus and SetLevel, the VVM project for the first time delivers a consistent methodological safety approach for automated driving in urban environments, allowing for industry-wide collaboration and added value. The project’s approach of scenario-based evidence of safety could, after regulatory approval, help to set global standards. VVM has created a reference system that is relevant for the future of the industry. It closes a methodological gap for practical validation and consolidates the pioneering role of German industry in automated driving among international competition.
VVM was funded by the German Federal Ministry for Economic Affairs and Climate Action with 26.7 million euros and initiated by the VDA flagship initiative “autonomous and connected driving” as a pre-competitive research project.