Here’s What You Need to Know About the Status of Autonomous Vehicles
Robotaxis are prowling the Las Vegas Strip at this year's Consumer Electronics Show, but widespread adoption of autonomous vehicle remains far from reality.
At this year's Consumer Electronics Show in Las Vegas, the popular ride-hailing service Lyft is offering robotaxis to weary pedestrians — self-driving BMWs powered by the company Aptiv (formerly Delphi), a major player in autonomous vehicle technology.
Aptiv and Lyft are running a fleet of eight vehicles at CES, and they can be digitally whistled up like any other ride, via the Lyft app. (For safety, a human driver is still behind the wheel, ready to take over.) Absent any crises, the Lyft cars will spend the week driving themselves and shuttling passengers to 20 different destinations along the Strip.
All indications are that the Lyft service is working perfectly fine as a rolling tech demo — and working spectacularly as a publicity stunt. But analysts and experts caution that the road to true self-driving vehicles remains a long one.
In an effort to cut through the hype, as 2018 gets underway, we sifted through the news, spoke to the experts and crunched some numbers. Here are eight things to know about where we're currently at with autonomous vehicle technology and the abiding dream of driverless cars.
There are many, many companies working on autonomous vehicle technology.
Hundreds of automakers, startups, and tech companies are developing self-driving technology to some extent, said Sam Abuelsamid, senior analyst at Navigant Research, a technology research firm in Detroit, Michigan. The major players in the US market include General Motors, Ford, Chrysler, Volkswagen, BMW, Honda, Nissan, Mercedes-Benz, and — of course — Tesla.
“On top of that, there are dozens of small startups that are doing testing across the US and around the world,” Abuelsamid told Seeker. Automakers are also partnering with ride-hailing services like Uber and Lyft and dozens of specialized technology companies.
Real-world tests on public roads are just getting started.
If it seems like there are a lot of public testing programs these days, that's because there are. Get used to seeing strange-looking cars with strange-looking logos navigating the streets and intersections of a city near you.
Uber has been typically aggressive in this area and was one of the first companies to do full-fledged testing in a major city when they deployed a fleet of self-driving cars in Pittsburgh in 2016. “They basically moved in, poached a significant portion of the research staff from Carnegie Mellon, and started testing,” Abuelsamid said.
Waymo, a subsidiary of Google's parent company Alphabet, also started early, testing in and around the San Francisco Bay Area. General Motors's recently acquired Cruise Automation division is now testing in San Francisco, using the all-electric Nissan Volt. Other busy testing areas include Phoenix, Arizona; Austin, Texas; Detroit, Michigan; and Kirkland, Washington.
California, a global hub for autonomous vehicle development, is probably the single busiest state for real-world testing. The California DMV has set up a useful public website for tracking which companies are negotiating state approval for testing on public roads.
You can expect more cities to get in on the action. According to the Initiative on Cities and Automated Vehicles, which is funded by Bloomberg Philanthropies, 37 US cities are hosting test programs or have committed to doing so in 2018.
The first wave of driverless cars will likely be introduced via ride-sharing services.
Virtually everyone in the business agrees that for Joe Commuter the first automated vehicles experience will come via ride-hailing services like Uber or Lyft — or whatever ride services emerge as industry leaders in the next few years.
“If we’re talking cars that are driverless from the start to endpoint of their journey, then ride-sharing services like Uber and Lyft will offer them at the front end of the shift to self-driving,” Pete Bigelow, technology and mobility editor with Car and Driver magazine, told Seeker.
Prognostication is usually a dubious practice with emerging technology, but Abuelsamid agrees that this particular prediction is a virtual lock.
“That's the way all of these cars are going to first come to market, as ride hailing vehicles,” he said. “In North America, Lyft and Uber are the two big players, but there are other big companies in China, India, and Europe.”
The very first driverless vehicles will be more like trains or buses, offering route transportation for groups of people at designated pickup and drop-off locations.
In fact, that's already happening in multiple test programs around the world — the Lyft service at this week's CES is basically just another shuttle going up and down the Strip.
“I think the first true iterations will be in autonomous shuttles — these 14-passenger minibuses that look like toasters,” Bigelow said. “The beauty of these is that they’re designed to drive slow, at 25 miles per hour or less to avoid federal motor vehicle safety standards. So, you’ll probably see these deployed either on private campuses like universities or giant office parks, or again, on fixed, defined routes that might run from one end of a popular downtown area to another.”
“These companies are already operating in some circumstances in Europe,” Bigelow said, “and we’re starting to see pilot projects in the United States.”
Down the line, you could have a lot of choices when you call up a driverless ride-hailing service.
Right now, Uber and Lyft are the dominant players in the US ride-sharing market, but you can expect niche players like Gett and Via to expand in coming years, Abuelsamid said. What's more, the very companies that are developing driverless technology are now very likely to introduce their own ride-hailing services as the technology evolves.
“The thing to keep in mind is that what Uber and Lyft and all these ride-hailing services provide — that logistics platform — is actually not that hard to replicate,” Abuelsamid said. “What we're actually going to see in the coming years is that all these companies that are developing automated vehicles, they're also going to be developing their own ride-hailing platforms. They'll deploy some vehicles through Uber and Lyft, but they're also going to develop their own services.”
The key to this development, Abuelsamid said, is that there is no significant “switching cost” if consumers decide to try a new ride-hailing service.
“If you switch from Android to iOS, it's a pain and you have to download everything all over,” he said. “Or if you switch social media platforms, you have to get all your friends to switch too, or it doesn't make sense.”
Not so with ride-hailing, he said.
“To switch ride-hailing services, there's no switching cost — you can go from Uber to Lyft in as long as it takes you to download the app,” Abuelsamid said. “There's nothing to keep you on any particular service. So, it's going to be a lot easier for other players to come into that market.”
If you want to own your own driverless car, be prepared to wait a while.
When we think of a full-fledged driverless car, that's what the industry calls Level 4 or Level 5 automation, Bigelow said. That's the scenario where you can climb into your car and ride under full autonomous control from your driveway to your destination.
“That’s not happening anytime soon,” Bigelow said. “Joe Average commuter isn’t going to a dealership to buy his Level 4 fully autonomous car until the late 2020s or 2030. We’re a long way away from anything approaching that scenario.”
Abuelsamid agrees. “Nobody is going to be selling Level 4 autonomous cars to consumers in the near future,” he said. “There are a whole lot of reasons why they're doing that, from a business standpoint and product liability concerns — that's a whole other conversation. The bottom line is that you're not going to be able to buy any of these cars from any of these companies anytime soon.”
When we do finally get autonomous vehicle on the roads, they won't look like they do now.
Many of the driverless test vehicles you see today in public test programs look almost comically unwieldy. You might see a stack of hardware on top of the cab, balanced like luggage from the “Vacation” movies, or odd sensors sticking out from various side panels.
That will change as the technology evolves, Abuelsamid said.
“Those big stacks on top of the vehicle, that's just what these things look like now, for development purposes,” he said. “But when they go to production, all the technology will be more tightly integrated into the overall design of the car.”
Finally, when it comes to discussions of driverless cars in the year 2018, it's useful to define your terms.
“Depending on the choice, the answers really change,” Bigelow said. “[Driverless technology] features like Cadillac’s Super Cruise and Tesla’s Autopilot are already here. Full self-driving all the time? A long way away.”
Bigelow suggests taking a different approach to the issue by asking different kinds of questions: When will driverless cars really change the way we get around? When will we feel the ramifications of driverless technology in society? When do they really reduce traffic fatalities and relieve congestion in cities?
“All those have different answers,” Bigelow said. “In the case of relieving congestion, maybe never. But I think the commonality would be — not for a long time. If you want to pin me down to a timeframe, I’d say late 2020s before we really get the first glimpses of the consequences.”