Virtual Radar: Simulation Technology for the Train Operations of the Future
Automated trains require highly precise sensors to reliably detect their surroundings – regardless of weather or lighting conditions. A newly developed virtual radar sensor model now makes it possible to simulate the behaviour of such sensors entirely within a digital environment. This allows complex scenarios of automated train operation to be realistically recreated and evaluated without the need to send physical trains onto the tracks.
The model was developed by DB InfraGO as part of the sector initiative Digitale Schiene Deutschland (DSD). It is based on the simulation environment NVIDIA Isaac Sim, which enables realistic modelling of radar waves, material properties and signal processing. In this way, the behaviour of a radar sensor on a train can be digitally tested – flexibly, cost-effectively and under reproducible conditions.
In an initial validation under laboratory conditions, the simulation results were compared with real radar data. The deviations were minimal – less than two centimetres in distance measurement and below one degree in angle detection. The virtual model thus behaves almost identically to a real sensor.
With the radar sensor model, DSD has created a foundation for developing sensor and object detection systems for automated trains efficiently and accurately – an important step towards the industrial development of series-ready sensor systems for digital railway applications.
A detailed technical article describing the background of the project was published in the October 2025 issue of SIGNAL + DRAHT.
