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.
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, 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 a bit 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 and VW invested a billion dollars into artificial intelligence outfit Argo AI, why General Motors bought a startup called Cruise, why Waymo has driven 20 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.
