PEGASUS – generally accepted approach for testing and approving highly automated driving

Automated driving functions are becoming established elements of road vehicles. The fundamental technical functionality has already been created and demonstrated in a multitude of situations. However, when in the future a machine controls the vehicle instead of a human, there must be a generally accepted approach related to the quality and performance of the automated vehicle’s behaviour.

The 17 PEGASUS project partners from the fields of industry and science are looking at key questions: What should an automated vehicle be able to do? How can we prove that it can do its task reliably? Introducing automated driving functions is practically impossible at the moment. Current tests and approvals are simply not designed to ensure automated vehicles for sale and public road use, because they are either too time and cost consuming or incomplete. The idea behind PEGASUS is to remedy this discrepancy. Thus, generally accepted criteria, benchmarks and elements of a tool chain are being established to ensure safe approval methods and processes for automated vehicles. Procedures which have so far been manufacturer-specific will be transferred to a generally accepted approach which assesses all vehicles according to the same criteria and standards. PEGASUS will therefore perform a key role in paving the way to the market introduction of automated vehicles.

The DLR’s role in PEGASUS is technical guidance and coordination: it is in charge of project coordination and running the project office and also works on content in sub-projects focussing on scenario analysis, quality benchmarks and testing. In cooperation with project partners, the DLR will develop the human-centric quality criteria and benchmarks for automated vehicles. This requires a precise understanding of human performance ability in road traffic. This is ascertained via assessments of prior Naturalistic Driving Studies and supplementary studies carried out in DLR driving simulators. By establishing the performance ability of machines, the DLR can apply its existing expertise from research into transition management, mode/situation awareness and human-machine cooperation and from work on balancing human-machine responsibility. These investigations into human performance ability are examined as to their transferability into generally applicable procedures for testing automated vehicles. The DLR will apply its knowledge in the fields of laboratory testing and live testing on closed-off and public roads and especially its AIM intelligent mobility application platform. AIM can be used to support and evaluate real-life test runs by using and expanding mobile parts of the AIM traffic and testing infrastructure  which already exists. A map system suitable for testing will also be established as a further development of the DLR projects SimWorld and SimWorld Urban. This will especially be used for laboratory and test bench work for model and simulation based tests. As a neutral research partner, there will be cooperative work with industry to evaluate the results as to their practicality and scalability.

Project name
PEGASUS – Project to establish generally accepted quality criteria, tools and methods in addition to scenarios and situations to approve highly automated vehicles functions
01/2016 – 06/2019
Federal Ministry for Economic Affairs and Energy
Project partners

Adam Opel AG
Audi AG
Automotive Distance Control Systems GmbH
BMW Group
Robert Bosch GmbH
Continental Teves AG & Co. oHG
Daimler AG
Deutsches Zentrum für Luft- und Raumfahrt
Forschungsgesellschaft Kraftfahrwesen mbH Aachen (fka)
iMAR Gesellschaft für inertiale Mess-, Automatisierungs- und Regelsysteme mbH
IPG Automotive GmbH
QTronic GmbH
TraceTronic GmbH
Technische Universität Darmstadt – Fachgebiet Fahrzeugtechnik
TÜV SÜD Auto Service GmbH
VIRES Simulationstechnologie GmbH
Volkswagen AG