The bus has to talk to the train

A conversation with Professor Anton Kummert, Dean of the School of Electrical, Information and Media Engineering at the University of Wuppertal (BUW).

Based on the five-level definition drawn up by SAE International, I’d say we are currently at Level 3 plus – and moving rapidly towards Level 4. To explain the definition, the highest level, Level 5, represents vehicle use that doesn’t require a driving licence or a steering wheel. Today, the driver must always be able to intervene while the vehicle is driving, for legal reasons alone. I think it’s conceivable that the first fully autonomous cars will be on the roads by 2025 – but they won’t be a mass phenomenon. What’s more, they will probably be available only in defined districts at first. That means that the vehicles will have learned the roads beforehand, so it won’t be so easy to transplant them from, say, Hamburg to Berlin, let alone from German roads to US or Chinese roads.

AI is vitally important to the mobility of tomorrow. As part of the project Bergisch.Smart_Mobility, which I’m involved with, we have defined various fields of action that can be fundamentally applied to any region of the earth. We’ve already mentioned one field of action: automated, connected driving. Here it’s important to look not just at the algorithms – the software – but to also consider the hardware. Because today it’s often still the case that prototypes are merely conventional cars absolutely packed full with sensors and a big computer in the boot. But of course, that’s not suitable for everyday life and hence far from being ready for mass production. That’s why in the project we are working to make the on-board electronics fit for the future – expanding the classic cable harness is simply not going to do it.

That’s why it’s so important to communicate the possibilities of AI honestly and in a way that is understandable. And, of course, there’s still lots to do. For instance, we need redundancies not just in the hardware, but in the software, too. Emergency strategies are essential. For instance, if scenario X or Y occurs, the AI must safely bring the car to a stop at the side of the road in a defined manner.

We talk about mobility on demand. In our project, this takes the form of shared taxis operated by the local public transport provider. The AI helps to plan the optimal routes: Not just collecting and transporting an individual person – you’ve also got to make sure that the passengers don’t have to wait too long for the shared taxi to arrive. It’s about multimodal transport concepts that need the individual modes to be very closely linked. To put it simply, the bus has to “talk” to the metro train, and the train has to talk to the shared taxi. Another field of action is infrastructure operators, such as local authorities or the Federal Highway Research Institute, the BASt.

By collecting and processing the almost infinite quantities of data from the transport systems. On that basis, the AI can intelligently plan traffic flows and control them in real time – for example using traffic lights or displays, such as those on motorways. Environmental aspects play a role here as well. To give another example, if sensors measure high particulate pollution on a particular road, the AI can ensure that heavy-goods traffic is intelligently rerouted.

Many thanks for this informative outlook.