Why not autonomous trains

Autonomous driving on the railways: What will happen to the train driver?

How do you rate the requirements for driverless driving in rail traffic?

Professor Schindler: Due to the signaling infrastructure, the monitoring of the routes and existing technologies such as LZB or ETCS, the basic requirements for driverless operation on the rails are already very good. However, the heterogeneous landscape of various operations control and safety systems still makes the train driver indispensable. In addition, the braking distances on trains are very long. An ICE with a speed of 300 km / h only comes to a standstill after three kilometers, while a regional train with 160 km / h is a good kilometer. When entering the train station, driving is done on sight, the view of the train driver is crucial here. In addition, unlike in the case of subways, we do not have an independent route in rail traffic, but rather openly accessible routes with interfaces to road traffic, for example at level crossings. After all, there are disruptions in the track, fallen trees as well as people. The responsiveness of a train driver is required here. Sensors can't do that, at least not yet, technology for cars is too short-sighted for trains.

What possible uses do you see for driverless trains?

Professor Schindler: Wherever rail vehicles run on closed tracks, driverless trains are already possible today, as is the case with subways or people movers. This also works in freight transport. We in Germany are researching the use of autonomous freight trains in marshalling yards. The marshalling yard then functions like a high-bay warehouse from industry, where pallets are automatically moved, temporarily stored over floors, pre-sorted and reloaded. To do this, the trains then need an automatic central buffer coupling. Many research projects are intertwined. I see good opportunities for the near future in the operation of non-fast moving trains, as they have shorter braking distances that are already covered by today's sensors. For example, the Australian mining company Rio Tinto has its own rail network in the outback, on which up to 40, exclusively driverless trains are in use. Siemens is currently testing an autonomous tram for passenger transport with the transport company in Potsdam. We ourselves at IFS are currently researching autonomous rail vehicles for railway lines in rural regions. Depending on the settlement and route, conventional trains run on such routes at most every 60 minutes. By using driverless railcars with the capacity of a standard regular bus, but with the comfort of a railroad, four journeys an hour could be offered instead of one. These rail buses would be slower, so that obstacle detection in connection with the braking distance problem could also be solved. So there are a lot of possibilities for autonomous driving on the railways, but the framework conditions have to be right. And there is still a lot to be done and researched.

So doesn't autonomous driving on the railway make the train driver superfluous for the time being?

Professor Schindler: Definitely not. Train drivers will certainly be needed for a few more decades. And there is also no economic pressure to replace train drivers with driverless trains. We are researching driverless rail transport precisely because there are too few train drivers. It is also a matter of making better use of the existing staff so that the offer on the railways remains stable and can grow. And even if trains can be controlled completely from the outside, train drivers are still needed. The job description will change, the job will remain. Engine drivers will then just work in the remote control or in troubleshooting.