How Automakers aim to gain from Quantum Computing
People who attend the Web Summit in Lisbon are familiar with how clogged the Portuguese city's streets get after thousands exit the event, where only one main traffic artery is available to ferry away visitors. This year, however, attendees will be able to download an app from Volkswagen Group and Canadian tech company D-Wave to predict the best route to any given destination. Real-time data is employed to mitigate the chaos, but the real purpose is to broaden awareness of the quantum computer that powers the calculations.
By optimizing the flow of people away from the event, the proprietary software helps improve urban productivity and reduce air pollution. "We would like to sell this to anyone interested in getting rid of congestion and reducing accidents," Volkswagen Group Principal Scientist Florian Neukart said. "We have already started to talk to different cities."
As innovation becomes increasingly software-driven, carmakers able to derive new revenue streams from endless lines of encoded ones and zeros will enjoy a competitive edge over rivals. Mining that data requires not just computing power, but speed, and that is the decisive advantage of these quantum machines. Because of their capacity to perform thousands of operations in parallel for a fraction of the energy, they can solve problems in seconds for which digital computers would need days, months or years and enormous amounts of costly electricity.
VW isn't the only automotive company tapping into quantum computing. Robert Bosch bought a stake in Harvard University spinoff Zapata Computing, which aims to be the leading developer of quantum algorithms. The world's largest auto parts supplier estimates market demand for such specialized software could reach into the billions within 10 years.
There are hurdles to overcome before "quantum supremacy," the point when quantum computers are undeniably more effective than their digital cousins, is reached. Yet automotive companies are already considering the broader implications this potentially game-changing technology poses for everything from mobility-as-a-service to artificial intelligence to optimizing the design of car components. "We see quantum computing as a strategic priority," said Joydip Ghosh, Ford Motor's expert on the subject. The inflection point for the auto industry may have come when VW Group Chief Information Officer Martin Hofmann teamed up with D-Wave more than two years ago to see whether the company's computers could offer VW practical help.
"When we started, the problems being tackled were typically academic, like prime number crunching," Hofmann recalled. "We wanted to learn what the opportunities and limitations were. Can we use quantum machines to model real-world problems?" Employing traffic data collected from thousands of taxis in Beijing, the duo could determine the most efficient route for 418 cars heading to the airport such that congestion would not even occur. In December 2017, researchers led by VW's Neukart published the groundbreaking study." This proved you could solve actual use cases such as traffic congestion if you can convert them into mathematical algorithms for mapping onto quantum hardware," Ford's Ghosh said.
Sven Gabor Janszky, a German trend researcher and futurologist, believes the VW experiment holds the key to eliminating infrastructure such as traffic lights one day. "All you need to do is transmit commands governing velocity and direction to the city's fleet, and there will be no such thing as a traffic jam anymore," Janszky said. Ford is currently working with NASA to examine how quantum computing can improve its business. One idea being explored is routing diesel delivery vehicles in such a way as to reduce harmful pollutants by ensuring their drive cycles are operating within an optimal range. Another idea is improving productivity in its manufacturing plants.
"There are many problems facing our industry that can be expressed in some form of the Travelling Salesman problem," Ghosh said. This well-known experiment, designed to find the shortest route between multiple destinations, along with similar variations, are ideal candidates for quantum computers. For example, automated guided vehicles at Japanese car parts supplier Denso operated 15 percent more efficiently with D-Wave's help. Industry researchers are only beginning to unlock the myriad potential this nascent technology provides. "Designing an engine block to minimize weight or finding the best way to configure a neural network: These are all problems that can be mapped onto a quantum chip," VW's Neukart said.
VW believes these advanced computers can also help them make batteries for electric vehicles up to 40 percent more powerful. It is this potential that interests Andreas Hintennach, who is Daimler's senior manager for battery research and technology. "Advances are occurring faster in the past 18 months than what we had even hoped," he said. "The field is attracting new university funding and an entire generation of young students. That means greater scientific exchange, more productive conferences and additional published research -- you cannot hope for anything better." Hintennach is part of an interdisciplinary team at Daimler working closely with partners IBM and Google, whose quantum machines can help him explore new active materials for a battery's electrolyte or electrodes. Currently this resembles a painstaking process of trial and error where new compounds are tested before placing samples under a scanning electron microscope to check for problems including degradation.
One approach to boost energy density is replacing the graphite in a cell's anode with lithium. Unfortunately, tiny grains of the metal are shed when charging or discharging, altering the texture of its surface -- a behavior called "lithium dendrite formation" that currently makes it unsuitable in combination with composite polymer electrolytes, for example. "If we could accurately predict how just four or five atoms in a functional group interact, it would already assist our research into lithium anode interfaces with solid electrolytes," Hintennach said. He believes he can more quickly envision how electrons from various materials, in particular electrolytes, may break or form chemical bonds with the help of a quantum computer.
While their very name sounds abstract, it stems from the time when science first turned its attention to the tiniest building blocks of matter and energy, initially referred to as quanta. Over a century ago, researchers including Albert Einstein discovered that objects on a subatomic scale do not behave according to the same Newtonian laws governing those visible to the naked eye. Thus began the study of these bizarre effects they called quantum mechanics. Typically, IT engineers treat these phenomena like unwanted gremlins, because they can introduce errors into computer calculations. By comparison, companies such as D-Wave harness their unique traits to model complex systems, achieving blindingly fast results that may not ultimately be 99.9 percent correct, but are close enough to suit their purpose. "A D-Wave quantum computer doesn't add, it doesn't subtract, it doesn't do things you are used to," said Bo Ewald, who until late March was head of the Canadian company. He is now CEO at U.S.-based startup ColdQuanta. "It's a fundamentally different way of thinking about computers."
Europe has fallen behind in quantum computing, however, so to address the deficit the EU Commission is financing a 10-year initiative that promotes research into quantum technologies to the tune of 1 billion euros. Germany's renowned laboratory in Jülich is tapping a portion of these funds to design and build a computer with up to 100 qubits in three years. With so many resources pouring into the field, it may even be only a matter of time before the technology becomes so widespread that an automotive company could purchase such a machine to guarantee unrestricted access. "People said at first employees would never have a desktop and now everyone has one," Daimler's Hintennach said. "I wouldn't rule out that we might have our own quantum computer one day."
Volkswagen is so convinced by the technology it has filed patents in the United States for three related applications, Hofmann said. The CIO imagines creating an "air traffic control" for urban cars he's calling quantum routing. Consultant Marcus Winkler believes pioneers such as VW Group will then be able to monetize their head start in this disruptive technology. The global head of Capgemini's automotive practice argues that quantum algorithms developed by scientists such as VW's Neukart and utilized by CIOs such as VW's Hofmann could power everything from mobility services to predictive maintenance.
"Hofmann is one of the CIOs who understands very well that new technologies have to be explored at a very early stage," Winkler said. "Quantum computing can really be a facilitator for technologies like machine learning and have a huge influence in many smart cities." The Lisbon trial, in other words, could just be the starting point of something much, much bigger.