Training Program for Control Centers in Autonomous Public Transportation

Autonomous shuttles are fundamentally transforming public transportation—and, at the same time, the demands placed on control centers. While the vehicles operate autonomously during regular service, control center staff remain the decisive authority whenever exceptional situations require a quick and safe response. Technical monitoring, situation assessment, and coordinated intervention are core tasks for which the legal framework—in particular §14 AFGBV—mandates structured training. In practice, however, this training quickly reaches its limits: autonomous vehicles in on-demand transport are only available to a limited extent for training purposes, real-world testing is costly and resource-intensive, and safety-critical scenarios cannot be specifically induced or systematically repeated during ongoing operations.

At Fraunhofer IVI, we are developing a simulation-based training environment that fully replicates real-world control center operations in autonomous public transport and makes safety-critical situations reproducible, teachable, and assessable. The environment is based on the open-source simulator CARLA and Unreal Engine and enables the digital reconstruction of real operational areas—including the faithful emulation of vehicle states, mode changes, and telemetry and error messages. Via a dedicated trainer interface, scenarios can be triggered specifically and repeated as often as desired under identical conditions—from vehicle malfunctions to medical emergencies on board. At the same time, various automation architectures can be integrated regardless of the manufacturer, allowing the platform to be flexibly tailored to different vehicle systems and control center environments.