Artificial Intelligence on Rails: ZF Makes Tram Stops Significantly Safer
- More non-automotive application opportunities for ZF ProAI, the central computer capable of deep learning
- Environment sensors recognize sources of danger, reducing reliance on the operator and increasing safety during operation
- Interfaces for drivelines and actuators prepare trains for highly automated functions
Berlin/Friedrichshafen. At InnoTrans 2018, ZF premiered the first potential use of ZF ProAI in rail vehicles. The central computer is capable of deep learning and, in connection with ZF environment sensors, can significantly reduce reliance on rail operators when starting off. The system recognizes and indicates potentially dangerous situations, such as when passengers are present inside the safety area. This significantly increases safety at stations and stops. As ZF ProAI already features the appropriate interfaces to drivelines and actuators, highly automated and interconnected functions can be achieved in rail vehicles.
The German Federal Statistical Office listed just under 1,300 accidents with physical injuries in German tram transport in 2016. Rail operators know: Fatal collisions are a significant factor, especially at urban stops, when pedestrians carelessly cross the rails or motorists do not notice a warning signal at the crossing. This is why ZF is helping to make rail traffic accident-free with its passive assistance system designed for urban rail.
Today, drivers have to check several screens and mirrors before starting off in order to ensure that no persons or objects are jeopardized when the vehicle starts to move. Here, a safety function that makes use of a smart algorithm can offer support. Using data from cameras, LIDAR and radars, it calculates a detailed image of the surroundings. This allows drivers to constantly keep an eye on their surroundings to a radius of six meters. Another benefit: The system automatically recognizes hazards, flags them on the screen and brings the obstacles to the train operator’s attention.
Benefit from the synergy effect
The foundation of this new safety function is ZF ProAI, a high-performance central computer that can process deep learning algorithms and thus make artificial intelligence possible. "Here, we benefit from the experiences gained in our automotive business," explained Lukas Hildebrand, head of the Automated Operation Off-Highway project house at ZF. "Our automotive colleagues already tested the connection between environment sensors and ZF ProAI extensively. This supports us greatly in the development of intelligent off-highway systems. In cities, rail vehicles travel at about the same speed as automobiles. What's more, they are subject to the same accident risks, especially due to pedestrians, cyclists and passenger cars."
High integration capability brings advantages for customers and operators
Thanks to its digital interfaces, ZF ProAI is already well prepared to control a variety of features. Moreover, it allows the integration of existing systems including those by other manufacturers. This would allow an automated exchange of information between vehicles and infrastructure – for an intelligent, safe urban traffic concept.