Automotive Industry Insights

KIsSME

Project duration: January 2021 – December 2023
Funded by the German Federal Ministry for Economic Affairs and Energy (BMWi)

In regard to the report and the study group results KIsSME is deemed relevant because of its close relation to open-road testing.

The KIsSME [1] (“Artificial intelligence for selective near-real-time recordings of scenario and maneuver data in testing highly automated vehicles”) project investigates how artificial intelligence can reduce the amount of data generated during the development of ADAS functionality.

Especially during the testing of highly automated vehicles, large amounts of data are generated. The KIsSME project aims to reduce this by developing an AI-based method that recognizes critical driving scenarios. Only if the current driving scenario is considered critical is the corresponding data recorded. This means recording only those data fragments during driving that add value for development and evaluation.

This approach will reduce and condense the accumulating data volumes to save storage space, electricity, and evaluation effort. The recorded data of relevance will be used off-board for rebuilding ASAM-standard-conform scenarios which can be used for the further development of automated driving functions, in terms of training, validation, and simulation.

The joint project is coordinated by AVL Deutschland GmbH, with Fraunhofer Institute for High-Speed Dynamics, Ernst-Mach-Institute, Forschungszentrum Informatik, Karlsruhe Institute of Technology, LiangDao GmbH, Mindmotiv GmbH, RA Consulting GmbH, and Robert Bosch GmbH participating as partners.

ASAM e.V. () as well as Cluster Electric Mobility South-West / e-mobil BW GmbH, State Agency for New Mobility Solutions and Automotive Baden-Württemberg are acting as associated partners.

[1] KIsSME stands for “Künstliche Intelligenz zur selektiven echt-zeitnahen Aufnahme von Szenarien- und Manöverdaten bei der Erprobung von hoch-automatisierten Fahrzeugen.”