As ADAS systems scale in complexity and deployment, the question is no longer whether they work, but how reliably they perform across the full range of real-world conditions. Traditional validation approaches, built around physical testing and controlled scenarios, are being pushed to their technical and financial limits by rising regulatory demands.
Jacob Perrin of dSPACE discusses how high-fidelity simulation to continuous, data-driven validation, is redefining how to build confidence in systems that must operate far beyond the lab or proving ground.
Written by:
Jacob Perrin
Autonomous Driving and Software Engineering
What needs to happen next to make ADAS safer and more reliable at scale?
The next phase of ADAS safety depends on reducing reliance on physical testing while also dramatically increasing test coverage. Regulations like FMVSS 127 raise the bar by requiring verified performance across a wide range of scenarios. Meeting those requirements through repeated provingground tests alone is not sustainable from a technical or financial perspective.
This challenge is especially apparent in the nighttime scenarios defined in FMVSS 127. Reproducing them requires controlled lighting, precise targets, and exact environmental conditions. These tests are expensive to schedule, difficult to repeat at scale, and frustrating for engineering teams who must staff them outside normal working hours.
What’s needed is a shift toward frontloading validation earlier in development by using highfidelity virtual testing that mirrors regulatory and realworld complexity. Scenariobased simulation enables engineers to perform tests that are impractical or unsafe to reproduce repeatedly in the real world. Effective use of simulation allows teams to iterate faster, detect issues earlier, and reserve physical testing for confirmation rather than discovery.
How can we ensure ADAS systems are truly improving safety in realworld driving, not just under ideal conditions?
Realworld safety depends on performance under variability: different road users, weather, lighting, road geometry, sensor degradation, and unpredictable human behavior. Passing a regulatory test or demonstrating performance in a controlled demo is no longer enough.
To close the gap between lab results and onroad reality, ADAS validation must be continuous and data-driven. This means grounding virtual scenarios in real traffic data, expanding coverage beyond requirements, and validating behavior across the full operational design domain. Coupling simulation models with software, hardware, and vehicleintheloop testing helps ensure consistency from algorithm design through production implementation.
Closing Remarks
As vehicles move further toward software-defined architectures, the way safety is validated must evolve alongside them.
Relying on physical testing alone is no longer viable at scale. The cost, time, and practical limitations of reproducing complex scenarios mean that traditional approaches cannot deliver the coverage required. Instead, the industry is shifting toward software-driven validation, where high-fidelity simulation and real-world data work together to expand test coverage and improve confidence in system performance.
This is why at AutoSens USA, experts including Jacob Perrin will be exploring the most critical industry trends, challenges, and advancements, that are shaping the future of ADAS and AVs.
Interested in In-Cabin monitoring technology?
With a pass to AutoSens USA, you’ll also get full access to our co-located sister event, InCabin. Take a look at the full agenda for InCabin here >>
Explore more interviews on the AutoSens YouTube Channel⬇
