HYPER-REALISTIC RADAR SIMULATION - ASCS Brochure #2025 - Magazine - Page 43
HYPER-REALISTIC RADAR SIMULATION
HYPER-REALISTIC
RADAR SIMULATION
Cut Costs, Save Time and Boost Trust in Sensor Design Interview with Dr.-Ing. Michael Stelzig, fiveD
You are always talking about hyper-realistic radar simulation. What exactly does fiveD develop?
We model how electromagnetic waves interact with
real-world environments — not just geometry, but also materials, motion, and multi-path effects. This enables engineers
to test sensors under highly realistic and repeatable conditions — long before physical prototypes exist. This approach
can be beneficial for numerous industries – from smart home
to autonomous driving.
How does AI relate to your simulation approach — and
where does it benefit from your data?
Our simulation is fully physics-based — we don’t use AI to
approximate radar behavior. However, AI plays a major role
around the simulation: Our high-fidelity, labeled radar data
enables better training and validation of perception systems
under realistic and reproducible conditions.
Why is it so difficult to simulate radar realistically?
Radar involves wave-based physics at very high frequencies.
Small details — such as surface roughness or material transitions — can significantly affect results. Capturing these
effects in dynamic, complex environments while keeping
simulations efficient is a major technical challenge that is
often oversimplified or misunderstood.
What’s your next major milestone — and what role does
your partner network play in reaching it?
We’re launching our first public demo version, which will be
made available to selected partners ahead of the official
release. Our network — including strong partners like Rohde
& Schwarz — plays a key role in this phase. Their early input
helps us refine the product and ensure it meets the needs of
both technical users and decision-makers. It’s all about building confidence and traction before a broader market entry.
What needs to change in the industry for realistic
simulation to become more widely adopted?
The biggest barrier isn’t technology — it’s confidence. Many
teams hesitate to invest in simulation for fear of wasting time
or budget on tools that may not deliver. What the industry
needs is more willingness to take calculated risks and trust in
high-quality virtual validation — especially now, with shift-left
becoming a key goal in the development cycle. With the right
models and methods, simulation can create enormous value
— but only if companies are bold enough to move beyond
conventional testing routines.
Dr.-Ing. Michael Stelzig is one of the co-founders of fiveD and is actively shaping the company’s
strategy and business development. Before co-founding fiveD, he obtained his PhD at the
Institute of Microwaves and Photonics (FAU Erlangen), focusing on radar system design for
extraterrestrial applications. Industry positions at Siemens and Mercedes-Benz complement
his experience at the intersection of high-frequency engineering and applied innovation.
www.fived.ai
43
