In the past five years, autonomous driving has gone from “maybe possible” to “definitely possible” to “inevitable” to “how did anyone ever think this wasn’t inevitable?” to "now commercially available." In December 2018, Waymo, the company that emerged from Google’s self-driving-car project, officially started its commercial self-driving-car service in the suburbs of Phoenix. The details of the program—it's available only to a few hundred vetted riders, and human safety operators will remain behind the wheel—may be underwhelming but don't erase its significance. People are now paying for robot rides.
And it's just a start. Waymo will expand the service's capability and availability over time. Meanwhile, its onetime monopoly has evaporated. Smaller startups like May Mobility and Drive.ai are running small-scale but revenue-generating shuttle services. Every significant automaker is pursuing the tech, eager to rebrand and rebuild itself as a “mobility provider” before the idea of car ownership goes kaput. Ride-hailing companies like Lyft and Uber are hustling to dismiss the profit-gobbling human drivers who now shuttle their users about. Tech giants like Apple, IBM, and Intel are looking to carve off their slice of the pie. Countless hungry startups have materialized to fill niches in a burgeoning ecosystem, focusing on laser sensors, compressing mapping data, setting up service centers, and more.
This 21st-century gold rush is motivated by the intertwined forces of opportunity and survival instinct. By one account, driverless tech will add $7 trillion to the global economy and save hundreds of thousands of lives in the next few decades. Simultaneously, it could devastate the auto industry and its associated gas stations, drive-thrus, taxi drivers, and truckers. Some people will prosper. Most will benefit. Many will be left behind.
It’s worth remembering that when automobiles first started rumbling down manure-clogged streets, people called them horseless carriages. The moniker made sense: Here were vehicles that did what carriages did, minus the hooves. By the time “car” caught on as a term, the invention had become something entirely new. Over a century, it reshaped how humanity moves and thus how (and where and with whom) humanity lives. This cycle has restarted, and the term “driverless car” will soon seem as anachronistic as “horseless carriage.” We don’t know how cars that don’t need human chauffeurs will mold society, but we can be sure a similar gear shift is on the way.
The First Self-Driving Cars
Just over a decade ago, the idea of being chauffeured around by a string of zeros and ones was ludicrous to pretty much everybody who wasn’t at an abandoned Air Force base outside Los Angeles, watching a dozen driverless cars glide through real traffic. That event was the Urban Challenge, the third and final competition for autonomous vehicles put on by Darpa, the Pentagon’s skunkworks arm.
At the time, America’s military-industrial complex had already thrown vast sums and years of research trying to make unmanned trucks. It had laid a foundation for this technology, but stalled when it came to making a vehicle that could drive at practical speeds, through all the hazards of the real world. So, Darpa figured, maybe someone else—someone outside the DOD’s standard roster of contractors, someone not tied to a list of detailed requirements but striving for a slightly crazy goal—could put it all together. It invited the whole world to build a vehicle that could drive across California’s Mojave Desert, and whoever’s robot did it the fastest would get a million-dollar prize.
Great for spotting things like lane lines on the highway, speed signs, and traffic lights. Some developers think that, with better machine vision, they can use cameras to identify everything they see and navigate accordingly.
The spinning thing you see on top of most self-driving cars is lidar (that’s “light detection and ranging”). It fires out millions of laser beams every second, measures how long they take to bounce back, and uses the data to build a 3D map that’s more precise than what radar offers and easier for a computer to understand than a 2D camera image. It’s also crazy expensive, hard to manufacture at scale, and nowhere near robust enough for a life of potholes and extreme temperatures. Good thing dozens of startups and tech giants are pouring millions of dollars into fixing all that.
At its simplest, this artificial intelligence tool trains computers to do things like detect lane lines and identify cyclists by showing them millions of examples of the subject at hand. Because the world is too complex to write a rule for every possible scenario, it’s key to have cars that can learn from experience and figure out how to navigate on their own.
Before a robocar takes to the streets, its parent company will use cameras and lidars to map its territory in extreme detail. That reference document helps the car verify its sensor readings, and it is key for any vehicle looking to know its own location, down to the centimeter—something standard GPS can’t offer.
A regular presence in cars since the late 1990s, radars bounce radio waves around to see their surrounding and are especially good at spotting big metallic objects—other vehicles. They’re cheap, reliable, and don’t sweat things like fog, rain, or snow.
The 2004 Grand Challenge was something of a mess. Each team grabbed some combination of the sensors and computers available at the time, wrote their own code, and welded their own hardware, looking for the right recipe that would take their vehicle across 142 miles of sand and dirt of the Mojave. The most successful vehicle went just seven miles. Most crashed, flipped, or rolled over within sight of the starting gate. But the race created a community of people—geeks, dreamers, and lots of students not yet jaded by commercial enterprise—who believed the robot drivers people had been craving for nearly forever were possible, and who were suddenly driven to make them real.
They came back for a follow-up race in 2005 and proved that making a car drive itself was indeed possible: Five vehicles finished the course. By the 2007 Urban Challenge, the vehicles were not just avoiding obstacles and sticking to trails but following traffic laws, merging, parking, even making safe, legal U-turns.
When Google launched its self-driving car project in 2009, it started by hiring a team of Darpa Challenge veterans. Within 18 months, they had built a system that could handle some of California’s toughest roads (including the famously winding block of San Francisco’s Lombard Street) with minimal human involvement. A few years later, Elon Musk announced Tesla would build a self-driving system into its cars. And the proliferation of ride-hailing services like Uber and Lyft weakened the link between being in a car and owning that car, helping set the stage for a day when actually driving that car falls away too. In 2015, Uber poached dozens of scientists from Carnegie Mellon University—a robotics and artificial intelligence powerhouse—to get its effort going.
After a few years, the technology reached a point where no automaker could ignore it. Companies like Ford, General Motors, Nissan, Tesla, Mercedes, and the rest started pouring billions into their own R&D. The tech giants followed, as did an armada of startups: Hundreds of small companies are now rushing to offer improved radars, cameras, lidars, maps, data management systems, and more to the big fish. The race is on.
The Future of Self-Driving Cars
Let’s start with the question you definitely want to ask: When will self-driving cars take over? Answer: wrong question. The autonomous vehicle is not a single device that someday will be ready and start shipping. It’s a system, a collection of inventions applied in a novel way. And, remember, the advance of the original car was constrained and shaped by forces like the growth of the road network and the availability of gasoline. The takeover of the self-driving car will depend on a new set of questions—the questions you should be asking.
When will self-driving technology be ready? That may, improbably, prove the easiest bit of making this real for the people whose lives it will affect. The hardware, to start, is mostly there. Radars are already cheap and robust enough to build into mass-market cars. Same goes for cameras, and the artificial intelligence that turns their 2D images into something a computer can understand is making impressive strides. Laser-shooting lidar is still pricey, but dozens of startups and major companies are racing to bring its cost to heel. Some have even figured out how to use their photons to detect the speed of the things around them, a potentially key capability. Chipmakers like Intel, Nvidia, and Qualcomm are pushing down power requirements for these rolling supercomputers, while companies like Tesla are making their own chips.
A Photographic History of Self-Driving Cars
The real job is to endlessly improve the software that interprets that sensor data and uses it to reason about how to move through the world. The key tool for doing that perception work—seeing the difference between a stray shopping cart and a person using a wheelchair, for example—is machine learning, which requires not just serious artificial intelligence chops but also gobs upon gobs of real-world examples to train the system. That’s why Ford invested a billion dollars into artificial intelligence outfit Argo AI, why General Motors bought a startup called Cruise, why Waymo has driven 10 million autonomous miles on public roads (and billions more in simulation). Safe driving requires more than just knowing that a person is over there; you also have to know that said person is riding a bicycle, how they’re likely to act, and how to respond. That’s hard for a robot, but these budding Terminators are getting better, fast.
OK, but are they getting better fast enough? In March 2018, a self-driving Uber Volvo XC90 operating in autonomous mode struck and killed a woman named Elaine Herzberg in Tempe, Arizona. The crash raised a number of suddenly pressing questions about testing autonomous vehicles on public roads. Is the tech actually ready? How should regulators handle this weird in-between moment, when the robots are good but not good enough? Should these vehicles really be testing on public roads? Right now, it's not clear how the incident will affect the development of self-driving vehicles in the US. But judging from Waymo's very cautious rollout of its tech in Arizona, it's safe to say the folks in this space have respect for just how hard this problem is to crack.
Meanwhile, a less capable version of the tech is already on the market. Cadillac Super Cruise, Nissan ProPilot Assist, and Tesla Autopilot all keep car in their lane and a safe distance from other cars, allowing the people behind the wheel to take their hands off the wheel. In November 2018, Tesla debuted a feature called Navigate on Autopilot, which gives its cars (including those already on the road, thanks to an over-the-air software update) the ability to change lanes to get around slower drivers or to leave the highway when it reaches its exit. Yet the human driver is required to keep paying attention to the road and remain ready to take control if needed. That's because these systems are not especially capable: They can't see things like traffic lights or stopped firetrucks. The problem is that humans are not especially well suited for serving as backups. Blame the vigilance decrement. And as these features proliferate, their shortcomings are making themselves clear. At least two Tesla drivers in the US have died using the system (one hit a truck in 2016, another hit a highway barrier this year), and the National Transportation Safety Board has criticized Tesla for making a system that's too easy to abuse. CEO Elon Musk has defended Autopilot as a life-saving feature, but even he doesn't use it properly, as he made clear during a 60 Minutes interview in December. Moreover, the main statistic he has used to defend the system doesn't hold up, a Tesla "safety report" offers little useful data, and it's not clear, anyway, how to produce more reliable numbers—or smarter systems.
Next question: Can we build and operate these things en masse? The huge automakers that build millions of cars a year rely on the complex, precise interaction of dozens or hundreds of companies, the folks who provide all the bits and bobs that go into a car, and the services to keep them running. They need dealers to sell the things, gas pumps or charging stations to fuel them, body shops to fix them, parking lots to store them. The folks who want to offer autonomous vehicles need to rethink interactions and processes built up over a century. Waymo has partnered with Avis to take care of its fleet of driverless minivans in Arizona, and it’s working with a startup called Trov to insure their passengers. GM is rejiggering one of its production plants to pump out Chevrolet Bolts without steering wheels or pedals. Lidar maker Velodyne opened a “megafactory” in San Jose where it says it could make a million units a year if it needed to. Federal regulators are considering ways to certify vehicles that don’t conform to safety standards written with human drivers in mind. Various would-be providers are drawing up plans for operations centers, where humans can keep track of their robofleets and cater to customers or cars in need. Legislators and public officials at all levels are racing to keep up and keep control of their streets. A bill that would set national standards for governing the robocars is working its way through the Senate. In Chandler, Arizona—home to Waymo's base in Arizona—the fire, police, and planning departments have hustled to prepare.
And it’s not if these things will be deployed, but how. To start, forget the idea of owning a fully self-driving vehicle. The idea of a car that can handle any situation, anywhere you want to go, is decades off. Instead, expect to see these robocars either debut as highway-bound trucks or in taxi-like fleets, operating in limited conditions and areas, so their operators can avoid particularly tricky intersections and make sure everything is mapped in excruciating detail. To take a ride, you’ll likely have to use predetermined pickup and dropoff points, so your car can always pull over safely and legally. Meanwhile, the people making these cars will be tackling knotty, practical questions. They’ll be figuring out how much to charge so they can recoup the R&D costs, but not so much to dissuade potential riders. They’ll wrangle with regulators and insurance companies, and what to do in the inevitable event of a crash that brings in the lawyers and legislators and safety advocates. And then, they’ll have to figure out how to expand—which is when the real competition begins. Uber and Ford and Waymo and GM may all start their services in different cities, but eventually they’ll start fighting for turf. You know how fiercely Uber and Lyft fight for market share today, tracking drivers, trying to undercut each other, and piling up promotions to bring in riders? Now imagine that same fight with several times more competitors.
Here’s the question everyone should really be asking: How will this technology change your life? Well, your ride to the airport will get cheaper and safer. Your pizza will show up in a human-free robot, no tipping required. Your highway commute will become less of a drag. You might get blasted with ads tailored not just to you but to where you are and where you're going at any given moment. But that’s the basic stuff, the horseless carriage.
The truth is, it’s hard to imagine what people will do once vehicles can move about on their own, and once these things are so efficient that the cost of transportation falls to something approaching zero. It’s easy to conjure up a dystopia, a world where robocars encourage sprawl, everyone lives 100 miles from their job, and sends their self-driving servants to do their errands and clog our streets. The optimists imagine a new kind of utopian city, where this technology not only eliminates crashes but integrates with existing public transit and remains affordable for all users. Like the internet, these vehicles will reflect some of our worse impulses, but also channel our best.
Waymo's So-Called Robo-Taxi Launch Reveals a Brutal Truth)
This was the moment many were waiting for, when Waymo, the company born as Google's Self-Driving Car Project and widely hailed as the industry frontrunner, would launch its robocars in a commercial taxi service. The reality was underwhelming: Only a subset of people already enrolled in Waymo's Early Rider program get to participate, and safety operators will remain behind the wheel for the time being. It's the best evidence yet that making the truly driverless—and truly safe—car is among the greatest technological challenges of our age.
A Not-So-Sexy Plan to Win at Self-Driving Cars)
While the big names in this space—Waymo, Ford, General Motors, Uber—are going for ubiquity, smaller players have already moved to carve out their own niches. May Mobility, for example, is running or will soon launch (human-supervised) robo-shuttles in Michigan, Ohio, and Rhode Island. “Our sales pitch is not that we are autonomous,” CEO Edwin Olson says. “It’s that we provide a better level of service and we’re solving real transportation problems.”
Burger King's 1-Cent Whopper Gives a Taste of the Robocar Future
Burger King's oddball fast food gimmick—drive to a McDonald's, get a coupon for a one-cent Burger King Whopper—won't seem so strange once robots have taken the wheel. Depending on what sort of service you take (and what you're willing to pay for it), you might be blasted with ads tailored to who you are, where you are, where you're going, and how you're feeling. Creepy, right?
How Do Self-Driving Cars See? (And How Do They See Me?)
Perception is the hardest part of the self-driving challenge, and the folks working on autonomy have three key tools to do it: radar, cameras, and lidar. Here's how each one of them works, and how the computer tackles the really tricky bit—using their data to tell what's what.
Robocars Could Add $7 Trillion to the Global Economy
Autonomous tech isn’t just poised to generate piles of cash for whoever can harness it—it will fundamentally change one in every nine American jobs. The good news is that this tech will also create new waves of jobs, much like the rise of the automobile helped yielded new forms of employment, like office park construction workers and pizza delivery drivers. If you want a job in this new, self-driven economy, your best bets are IT and data crunching.
As Self-Driving Cars Approach, the Auto Industry Rushes to Rebuild
Automotive supply chains span tiers of companies scattered around the globe. Automakers are constantly seeking to reduce cost and complexity—and increase profit—by joining forces with the folks whose strengths match their weaknesses. Now the industry is building the supply chains and partnerships that well help it push into a new self-driving age, one that demands manufacturing expertise, artificial intelligence know-how, mapping skills, piles of cash, and more.
Mercedes-Benz's Plan for Surviving the Auto Revolution
If you're supposed to be leading one of the world's biggest companies through a maelstrom of change, it probably helps to feel a bit like a superhero. WIRED sat down with Wilko Stark, who's running Mercedes-Benz's future strategy, to talk about the rise of "mobility," a diversifying auto industry, the role of electric cars, and why AMGs aren't going anywhere.
For a Much-Needed Win, Self-Driving Cars Should Aim Lower
After a self-driving Uber killed pedestrian Elaine Herzberg in Arizona in March, the technology's defenders fell back on their usual arguments: Nearly 40,000 people die on US roads every year, and human error causes more than 90 percent of crashes. But no matter how quickly self-driving proliferates, it will be a very long time before the robots can put a serious dent in those numbers, and convince everyday folks that they're better off with less flesh behind the wheel. So if the tech companies want to win people over, they should aim for discrete, tangible goals along the way—maybe launch a service to take teens home from parties, or get bedroom community commuters to the train in the morning.
The Very Human Problem Blocking the Path to Self-Driving Cars
Maybe you’ve heard about Tesla’s Autopilot, Cadillac’s Super Cruise, or Nissan’s ProPilot Assist—systems in otherwise conventional cars that let the machine drive itself in limited situations, with human supervision. Sounds great, right? Yes, until you hear about the handoff problem. Making a semiautonomous car means designing a vehicle that doesn’t just drive itself but knows what its human is up to—and how to get them to take the wheel when needed.
Preparing Self-Driving Cars for the World’s Most Chaotic Cities
Robots love rules. Stop here. Yield there. Go this speed. But navigating a human-filled world means knowing how to bend those dicta, how to negotiate with other drivers, pedestrians, cyclists, and others. That’s especially true in cities in developing countries, where traffic laws don’t hold much sway. So how can you possible make a car that can handle driving in India, or Lebanon, or Vietnam? It’s an important problem to solve: Chaotic cities are where car crashes claim the most lives and where safer robots could make the biggest difference.
Lawyers, Not Ethicists, Will Solve the Robocar ‘Trolley Problem’
Giving machines the ability to decide who to kill is a staple of dystopian science fiction, and explains why three out of four American drivers say they are afraid of self-driving cars. But worries over the “trolley problem”—in a situation where a crash is unavoidable, how does the robot decide whom to hit?—might have already been solved. One Stanford researcher says the law has preempted the problem, because the companies building these cars will be “less concerned with esoteric questions of right and wrong than with concrete questions of predictive legal liability.” Meaning, lawyers and lawmakers will sort things out.
What Does Tesla’s Automated Semi Mean for Truckers?
Elon Musk is building a truck, and like other Tesla vehicles this semi will be able to handle itself on the highway. He’s not the only one trying to throw the driver out of the cab—and threaten 3 million solid middle-class American jobs in the process. So what does the future look like for truck drivers? That kind of depends on how you define trucking. Because autonomous big rigs aren't going to be 100 percent autonomous, at least not in the near or medium future.
Google’s Self-Driving Car Company Is Finally Here
Seven years after kickstarting the self-driving industry, Google’s self-driving car project became its own company under the Alphabet umbrella: Waymo. Now the company is rushing toward commercializing its miracle movers. "We're a self-driving car company with a mission to make it safe and easy for people and things to get around," says CEO John Krafcik. What that means, exactly, is still an open question—perhaps as much for Waymo as for the rest of us.
The End of Waymo v. Uber Marks a New Era for Self-Driving Cars: Reality
When Waymo and Uber settled their blockbuster trade secrets lawsuit (for a cool $254 million in Uber equity), it sent a clear signal across the still burgeoning industry: Time to grow up. The sector looks way different than it did when a Google engineer named Anthony Levandowski left for its grand rival, Uber, allegedly taking lidar secrets with him. There are more competitors now, and much, much more money. If the whole affair teaches us anything, it's that autonomy is a real business now. Act accordingly.
Uber's Self-Driving Car Just Killed Somebody. Now What?
This is what the self-driving car community feared all along: the first death associated with an autonomous vehicle. After a self-driving Uber (with a safety driver behind the wheel) struck and killed a woman in Tempe, Arizona, we were left with only questions. Is the tech really ready for prime time? Is the safety driver model broken? How will regulators respond?
Plus! Waymo's robotrucks and more WIRED self-driving car coverage.
Last updated December 13, 2018
Enjoyed this deep dive? Check out more WIRED Guides.