The Future of Autonomous Vehicles Is Shared
A dispatch from the 2017 Consumer Electronics Show, where robo-cars were on full display—and predictions about how we'll use them were flying.
At the 1939 New York World's Fair, General Motors dazzled tens of millions of attendees with a 36,000-square-foot scale model of the future, in which a vast network of superhighways stitched together downtowns and suburbs. I HAVE SEEN THE FUTURE, declared the souvenir buttons visitors wore home.
Thirty-six hours chasing self-driving cars through the sprawling exhibition halls and demo lots of the 2017 Consumer Electronics Show made me feel the same way.
What I can say is that the future of mobility actually looks fairly utilitarian. For one, most autonomous vehicles will not be marketed as something you'd personally own—at least not in the near future.
That is not exactly a secret. In August, Ford promised to roll out a "level four" autonomous vehicle—meaning no steering wheel, brakes, or gas pedal— designed for "high-volume" commercial uses like ride-hailing or package delivery by 2021. Looking at the self-driving models promised by other manufacturers, you get the sense they're being designed for similarly practical purposes.
General Motors, which already furnishes ride-hailing vehicles to its partner Lyft, is developing fully autonomous Chevy Bolts, its no-frills, all-electric subcompact. Alphabet's self-driving tech company, Waymo, is doubling down with a Chrysler minivan model—which would seem a good fit for the tech giant's nascent ride-hailing service.
But hearing from carmakers and tech insiders, and actually experiencing road-ready, shared autonomous vehicles on the move at CES, really drove the point home. On Wednesday, I rode the length of the Las Vegas Convention Center demonstration lot in the pod-like Arma, a friendly-looking self-driving shuttle from the French manufacturer Navya (check out my ride, in the video below). With 15 seats and no steering wheel, it's like a tram car detached from the track —except the Arma braked for people crossing its path, and knew to carefully avoid a semicircle of traffic cones staffers had set up for the demo.
With a top speed of about 28 miles per hour, the Arma is designed for use as a fixed-route bus or on-demand ride service in relatively small geographic areas, like a university campus or an industrial park. There are already 30 of them deployed around the world, with hundreds more coming online in the next year, according to Henri Coron, Navya's vice president of sales.
"Efficient, useful, slow and unexciting, it's much like what we can expect when self-driving cars hit the roads en masse," the tech journalist Loz Blain wrote of the Arma in 2016. (Personally, I couldn't rip the smile off my face as it hummed along its little obstacle course—but I guess the novelty will fade.)
Up until fairly recently, there wasn't clear consensus about whether self-driving cars would be for personal or shared use. Now there is. The prohibitive cost of self-driving cars is a huge part of the reason why AVs are likely to be shared, at least initially, but there are others, according to Jim McBride, a technical leader of autonomous vehicle development at Ford. He and I sat between a Chariot minibus (Ford's recent micro-transit acquisition) and a semi-autonomous hybrid Fusion, both parked on display at the automaker's massive convention exhibit.
McBride said his company figures it'll have more direct control over early AVs if they're designed as commercial fleets: "Normally, when you sell a car to a customer, you might never see them again," he said. But with a commercial fleet, "you could make sure that you really understand where they're being driven, that they're operating in appropriate weather for their state-of-art sensors, that the vehicles coming back every day, and that they're being properly maintained and serviced."
Carmakers keeping a close watch on shared or commercial fleets should help the cars' safety specs improve. Such a strategy could also help expose a still-wary public to the flavor of a robot's driving skills, McBride said. Given the opportunity to see or even take a ride in one of these cars—and feel the full-bore braking power in a confrontation with a pedestrian or another car—folks might get over that trust barrier.
What industry insiders do disagree on, however, is when (if ever) widespread autonomous vehicle adoption is likely to happen, and what effect it'll have on personal car ownership. McBride suggested that down the road, as the price of the technology comes down and acceptance becomes more widespread, Ford could start to market AVs to individuals.
Jay Ellis, the co-founder of the University of Michigan's TechLab at MCity, a transportation technology proving ground, said at a Wednesday morning panel that the transition to AVs—i.e., people giving up personal car ownership—would happen "more slowly than we think." Paul DeLong, the president and CEO of Car2Go North America, said that even as Americans start to adopt shared mobility services, they will still own and operate personal "primary" cars for some time. "People still enjoy driving," he said.
Others thought differently. "The next generation will have no problem getting in these vehicles," said Karen Francis, the board director and strategic advisor for a slew of automotive tech companies, including AutoNation, Nauto, and Telenav. David Baga, the chief business officer of Lyft, thinks people would be chucking out their car keys in serious short order. No doubt, the futures that these leaders predict are the ones that best serve their respective business models. But it's still a fact: The timeline, and eventual implications, of shared AV adoption are not clear.
The government's participation will help answer those uncertainties. Francis made an interesting semantic observation I'd never heard before: "Autonomous vehicle" is a misnomer, she said, because for self-driving cars to fully realize their safety and efficiency potential, a whole ecosystem of "smart" infrastructure must accompany them. Think street lights, roads, curbs, and parking spots equipped with sensors and special markings that "talk" to the cars. "These vehicles don't operate 'autonomously.' They have to operate within a larger connected world," she said.
Who's going to build that world? The public sector, of course—just as the government built out the autopia of the future that GM so thrillingly depicted back in 1939. The contours of urban roadways will eventually make room for self-driving vehicles. How quickly cities do this, and how well new roads facilitate ride-sharing—think infrastructure like passenger loading docks and off-street car storage—will determine that future.