XRAISE Project to Be Presented at InnoTrans 2026

Dr. Carsten Knoll and Dr. Sascha Weber from TU Dresden were among the 3,190 exhibitors at InnoTrans 2026. At this international trade fair for rail and transportation technology, 175,591 visitors gathered at the Berlin Exhibition Grounds from September 22 to 25, 2026. The interdisciplinary team from TU Dresden presented the XRAISE project : Safety Argumentation Using Explainable AI (xAI), funded by the German Center for Rail Transport Research at the Federal Railway Authority. The project was also presented to Ulrich Lange, Parliamentary State Secretary at the Federal Ministry of Transport.

XAI makes AI decisions traceable. In automated driving within open rail transport systems, an image-processing AI system could be part of the track monitoring system. However, there is currently no generally accepted approach for verifying the safety of deep neural networks. Conventional software standards such as EN 50657 are not directly applicable, as their decision-making processes are less transparent. Explainable AI (XAI) methods can create transparency here by showing which image regions and concepts influence the AI’s decision.

© TUD