SAFER’s project portfolio has been enriched with yet another project, EVIDENT - Enabling VIrtual valiDation and vErificatioN for ADAS and AD features.
In order for new functions in vehicles, which are intended to help the driver in various situations, to actually contribute to increased safety on the roads, testing and verification is a very important ingredient in the product development. As the systems become more and more complex, the testing also becomes more and more demanding. Here, a recently started project within the SAFER platform can contribute to increased road safety by addressing challenges to find enough necessary edge cases and feeding a suitable amount of tricky situations into verification and validation.
This project will explore strategies which trades degree of fidelity (how realistic the testing environment is) in a quantifiable way to complement or replace recordings from real traffic situations. This data could be captured by real sensors or by simulative components to turn retroactive validation and verification (V&V) strategies, which is driven by real-world recordings only, into a proactive V&V strategy. This would allow for a systematic testing of an adjusted software feature before it is deployed.
A successful strategy would allow all involved actors along the value chain of validating and verifying software-driven vehicle features, to more efficiently allocate the right testing resources, in the right quality, and with the expected degree of fidelity at the most pressing challenges driven by information from the field. An expected effect is that vehicle industry increases their testing efforts with regards to ADAS and AD testing, without the same need for expanding track and open road testing.
Vinnova is the financier and the total budget is about 15 msek. The SAFER partners involved in the project are AstaZero, Asymptotic, Chalmers, CEVT, Gothenburg University, RISE, Veoneer and VTI. Viktor Johansson, AstaZero, is the project leader. The researchers are planning to finalise the project mid 2024. SAFER's research area Systems for accident prevention and automated driving will host the project.
The project contributes to an important goal in SAFER's strategic research agenda; to support the development of methodologies to verify and validate assisted and automated systems.