Swiss Federal Railways uses AI for track inspections

Swiss Federal Railways (SBB) is utilizing artificial intelligence and specially equipped diagnostic trains to inspect approximately 2,000km of railway track every month. Unlike traditional manual inspections conducted at walking speeds, these diagnostic trains travel at speeds of up to 120kph, using cameras and track-geometry measuring devices to film and measure the rails.

AI systems analyze the captured images and data to identify damage and defects. While SBB began experimenting with this technology a decade ago, early attempts faced challenges with high error rates. Recent advancements in AI have improved the accuracy and reliability of the system. Currently, the two diagnostic trains in operation detect up to 1,500 defects monthly, the majority of which are minor.

Despite the technological integration, human inspectors remain essential. The AI is specifically trained to recognize certain defects in particular track sections; it may not detect issues outside those parameters, such as damage to overhead-line masts. Consequently, human oversight is still required to identify unforeseen problems and to perform necessary repair work.