What challenges do companies face when testing complex, safety-critical systems, and how does fortiss help address them?
Many companies face significant challenges when it comes to testing advanced, safety-critical systems—from increasing system complexity to meeting stringent regulatory requirements.
Often, automated methods or scalable tools to efficiently conduct safety testing are lacking. We work closely with our partners to bring advanced, research-based solutions into practical application. This includes providing automated, scenario-based testing tools, supporting architectural decision-making, and integrating our methods into existing development pipelines.
We also assist with certification processes and securing public funding to help reduce risks and costs. The result is a smoother path to safe, certifiable products.
How does aligning safety practices with industry standards benefit companies, and how has fortiss applied its safety methods in real-world projects to achieve these benefits?
Aligning safety assurance practices with recognized standards enables companies to bring complex systems to market more efficiently and with greater reliability. It ensures that safety-critical features meet certification requirements while keeping development agile and cost-effective.
At fortiss, we support this by combining solid research with real-world applications. For instance, we created OpenSBT, a scenario-based testing framework developed in collaboration with DENSO to evaluate functions like automated emergency braking. Since then, it has been extended to cover additional autonomous features and systems, such as lane-keeping and drones. This is one of the ways we help our partners scale their testing processes while staying aligned with safety standards.
AI in vehicles is becoming increasingly complex and evolving. How does fortiss ensure that these systems remain safe and meet certification standards?
As AI-powered vehicle systems become more complex, their behavior grows increasingly difficult to predict and validate using conventional testing approaches.
At fortiss, we address this challenge by integrating formal verification techniques with automated, scenario-based testing to continuously assess safety-critical components. This combination enables early detection of potential issues, even as systems evolve.
By aligning our methods with standards such as ISO 26262, we ensure that they are certification-ready and applicable in real-world development settings. Our ultimate goal is to deliver trustworthy, evidence-based assurance even within fast-paced, software-driven development cycles.
How does fortiss ensure that different test tools in the field of automated driving work better together, and how do industry partners benefit from this?
In the field of automated driving, companies frequently rely on multiple simulators that often lack seamless interoperability, leading to inconsistent or unreliable test outcomes.
At fortiss, we’re addressing this by developing cross-simulation and ensemble-based testing methods that integrate simulators with varying levels of detail.
These approaches enhance the reproducibility and reliability of testing processes, enabling smoother collaboration across different tools and platforms. Ultimately, this helps our partners develop more dependable systems and uphold stringent safety standards, even within complex and diverse tool chains.
How does fortiss use generative AI to improve the testing of autonomous vehicles?
We’re exploring how generative AI especially diffusion models can be used for real-time simulation in testing. Traditionally, diffusion models are used offline to generate training data. At fortiss, we go further: by combining generative AI with knowledge distillation techniques. So we can simulate complex driving scenes in real time.
This makes it possible to test autonomous systems under a wide range of realistic conditions, quickly and efficiently. It greatly enhances scalability because companies can run more tests earlier in development, in a controlled and repeatable environment.
