Weather testing framework for ADAS safety validation

The WeatherSafe Project has established a technical framework for evaluating advanced driver assistance systems (ADAS) and automated vehicles under severe weather conditions. The initiative focuses on improving system robustness and road safety by combining accident analysis, sensor evaluation, and test methodology development for low-visibility and reduced-grip scenarios.

Scenario-based risk modeling from real-world data
The project identified six representative accident scenarios based on real-world crash analysis, with particular emphasis on rain, solar glare, and road surface friction. These scenarios serve as standardized reference cases for validating ADAS performance and assessing risk within the automotive data ecosystem.

By grounding testing methodologies in observed accident conditions, the framework enables more realistic evaluation of system behavior in environments where sensor performance and vehicle dynamics are degraded.

Sensor performance in adverse weather conditions
Extensive testing over a three-year period enabled detailed characterization of how environmental factors affect onboard sensor systems, including cameras, LiDAR, and radar. The results highlight the impact of reduced visibility, precipitation, and light interference on perception accuracy.

These findings provide a basis for improving sensor fusion strategies and system calibration, particularly in conditions where individual sensing modalities exhibit reduced reliability.

Role of tire–road friction in system response
The project also quantified the influence of tire–road friction on vehicle safety functions. Accurate estimation of friction coefficients allows ADAS to better anticipate loss-of-control scenarios and adjust braking, steering, and stability functions accordingly.

This aspect is critical for enhancing system performance in wet or low-grip conditions, where traditional models may not sufficiently reflect real-world variability.

Virtual sensor modeling and hybrid validation
A key technical outcome is the use of virtual sensor modeling to simulate complex environmental conditions. These models are combined with laboratory testing and open-road validation to create a hybrid evaluation approach.

This methodology enables scalable testing of edge cases while maintaining correlation with real-world performance, supporting more robust validation of automated driving functions.

Contribution to standards and assessment protocols
The project defined five technical guidelines for evaluating safety systems under adverse weather conditions. These recommendations support the evolution of testing frameworks and contribute to the integration of weather-related parameters into regulatory and assessment standards.

The results are intended to inform future developments in ISO standards and consumer safety protocols such as those defined by Euro NCAP, particularly in relation to automated driving validation.

Experimental demonstrations and industry collaboration
The project included practical demonstrations, such as automatic emergency braking (AEB) tests conducted in fog and low-visibility conditions by Aptiv. Comparative evaluations of spray generation systems developed by ASTAZERO and MESSRING enabled assessment of reproducible test conditions for wet-weather scenarios.

These activities were complemented by technical roundtables involving research institutions, industry stakeholders, and assessment bodies, supporting alignment on future validation requirements.

Outlook for regulatory integration
The project outcomes provide a structured basis for incorporating weather conditions into safety validation frameworks. By combining data-driven analysis, simulation, and physical testing, the approach supports more consistent and measurable evaluation of ADAS performance.

The initiative is expected to inform future collaborative programs and industrial implementations, contributing to the development of safety standards that reflect real-world operating conditions more accurately.

Edited by Maria Brueva, Induportals editor – adapted by AI.

www.utac.com