Cooperation with Mittweida University of Applied Sciences and IAV GmbH: Development of a research environment that makes AI directly tangible in everyday life
As part of the AI in the Loop project, the DLR Institute for AI Safety and Security is working alongside Mittweida University of Applied Sciences (AI) and IAV GmbH to develop an infrastructure that will enable AI to learn, be retrained and monitored continuously during use. Providing containerised AI models on real systems allows for direct testing of practical suitability and continuous improvement. The findings and design principles resulting from this collaboration will later be generalised for use in safety-critical use cases. Mittweida University of Applied Sciences is developing the robots' driving functions and implementing the AI in the Loop process. IAV is responsible for the IoT platform that connects and controls the systems.
Contribution DLR-Institute for AI Safety and Security
The Institute for AI Safety and Security is coordinating the project, providing a cloud-based infrastructure, and offering a real training laboratory equipped with self-driving telepresence robots at its Sankt Augustin site for testing new learning methods. During ongoing operations, the AI models are flexibly controlled, updated and tested in real-world environments. The AI models are also evaluated in terms of their reliability during testing. Safety is then verified by taking into account the specific parameters of the operational design domain.
Institutes and facilities involved (DLR & external)
