Quantum computing could one day provide the huge boost in processing power that is needed to make future autonomous cars reliable, safe and secure. Volkswagen CIO Martin Hofmann says the technology is moving closer to commercialization.
automotiveIT: Mr Hofmann, please explain where Volkswagen is with ist quantum-computing research?
Martin Hofmann: Until now, quantum computing has been a topic for the scientists. But the hardware has reached a stage where it becomes interesting to start developing use cases. That’s why we want to find out whether we can build relevant applications using such systems. We’re asking whether specific problems and issues can be addressed with a quantum computer. VW researchers recently worked on an optimization problem with a quantum computer. The project involved publicly available driving data from taxis in Beijing. Our IT experts managed – after the fact – to use a quantum computer from D-Wave to create the most optimal routing for a few hundred taxis, so that they could all have reached their destinations without being stuck in traffic. We were trying to prove that a particular traffic optimization problem could be addressed with a quantum computer.
You have been cooperating with Canadian quantum computing specialist D-Wave for quite a while and with Google since last autumn. How do these cooperations work?
You access quantum computers remotely because they are located at the manufacturer. Google and D-Wave have the hardware and the knowhow to operate these computers, but we have the real-life test projects and applications for which we program the algorithms. This tie-in with the real world is interesting for the manufacturers.
Can you tell us a bit more about the use cases?
In addition to the traffic optimization I mentioned, we are looking at battery simulations and machine learning for the environment recognition of autonomous vehicles. Battery chemistry, for example, is very important in the charging and discharging cycles of electric vehicles. On a microscopic level, what happens there is so complex that we cannot today create a good-enough simulation. There’s a massive shortage of processing power, which is why the experts have to build physical prototypes. That costs time and money. With quantum computers, you will one day be able to do away with the prototype because the battery chemistry simulation is realistic enough. VW wants to build up expertise for this at an early stage.
What’s it like for a developer who has grown up in the binary age to work on a quantum computer?
Because a quantum computer works completely differently, you have to learn first to appropriately describe the problem. There are no experts for this today, so developers have to acquire this knowledge themselves. At Volkswagen in Munich and San Francisco we have around 80 people working on such issues; IT specialists, physicists and artificial-intelligence people. You have to see working on a quantum computer like working with assemblers in the 1970s. It’s real pioneering work that are our specialists are doing. Everything is machine-focused and directly on the processor – at least if you want to go deeper. But we do now have the first graphical programming interfaces and software libraries.
Interview by Michael Vogel