utonomous vehicles, alternative drive systems, soft-
ware-defined applications – the automotive industry
is one big playground for creative minds. But the task
is by no means child’s play. Developing alternative propulsion
systems requires expensive virtual simulations or complex
hardware. AI and digital solutions from autonomous vehicles
mean high computing power and huge volumes of data. This
is simply too expensive for small or otherwise specialized
companies.
Corporate partnerships can make an important contribution
to overcoming the numerous challenges through sharing
resources and test facilities and searching for market appli-
cations for innovations together. Therefore, in order to tap the
full potential of co-creation, AVL launched Creators Expedition
back in 2017. Here, AVL experts and business departments
come together with innovative partners from the fields of elec-
tromobility and alternative fuels, automated and connected
mobility, AI, and big data, with the goal of jointly pioneering the
development of innovations. This first requires a clear shared
vision and clear objectives to ensure that everyone is on the
same page and sees a clear benefit for themselves. Then we
can go on a joint “expedition” and reach out to new customers
and markets.
1. Co-Creation With Startups
We primarily consider early-stage companies that have a
strong team, a viable product, and some experience. With
our expertise and resources, we give them the opportunity to
focus fully on developing their product. If the new technology
is mature but the market is not yet ready, we also help them
find more immediate applications.
Our Answer:
Creators Expedition
Objective: Substantial reduction of development times and budgets for early
design and simulation variants, as well as lower project risk due to greater
variance of possible design variants
Solution: 3D deep learning approach to bypass classical FE simulation loops
in early development phases
“Conventional approaches require several dozen FEM simulations to be run
through for various design variants of components throughout the entire
development process. This consumes a lot of resources and time and is
difficult to reconcile with ever shorter development cycles. With our techno-
logy, efficient optimization is possible at an early stage, guaranteeing high
performance and the reduction of expensive errors that occur later. For the
chosen use case, the proof-of-concept project with AVL clearly demonstra-
ted that our deep learning approach saves a massive amount of time even
with relatively little training effort required for the models. At the same time,
the expectations regarding the precision of the results are exceeded signifi-
cantly. We can now build on these promising results in our collaboration with
AVL in other fields.” Kosmas Kritikos, Commercial Director, Neural Concept
AVL and Neural Concept
AI as a Turbo Booster for Design,
Simulation, and Prototyping
SPECIAL EDITION