AI4SDV – AI for Software-Defined Vehicles
Today, around 80 percent of automotive innovations originate in software development. However, while software complexity has quadrupled over the past decade, development productivity has stagnated, compounded by a severe shortage of skilled professionals. The EU project AI4SDV addresses this bottleneck, using Artificial Intelligence to radically accelerate and streamline the development of Software Defined Vehicles (SDVs).
AI as an Engineering Tool
AI4SDV aims to relieve engineers of time-consuming routine tasks, freeing up capacity for creative, value-adding activities. The project combines established systems engineering practices (such as model-based development) with cutting-edge AI technologies (such as Generative AI). It develops modular tools for the entire lifecyle, from automated code and test generation to early error detection in system architecture.
Strong Consortium & Robust Infrastructure
The project unites over 50 partners from industry and academia, including Mercedes-Benz Tech Innovation, Valeo, Infineon, dSPACE, Fraunhofer, and TUM.
HLRS contributes its expertise in high-performance computing, providing the massive processing power required to train and run large AI models via its "HammerHAI" AI factory. Additionally, HLRS focuses on optimizing the scalability, efficiency, and security of these AI workloads.
Expected Impact
The ultimate goal is a production-ready, open AI platform that shortens development cycles, reduces costs, and improves software quality. Thanks to an open-source approach, these results will benefit the entire European automotive industry.
