Once the stuff of science fiction, autonomous vehicles are almost here. The promise is greater convenience, fewer accidents and improved traffic flow, but there are still some bumps in the road.
There’s something strange about sitting behind a steering wheel with your hands neatly folded in your lap—especially when the little Nissan Leaf suddenly surges to life without even the tap of a toe on the throttle. My copilot seems entirely relaxed, however, so I try to just smile and take in the scenery at the old Miramar Marine Corps aviation base near San Diego as the battery car bobs and weaves its way down the road. That’s when I see the traffic zigging and zagging through the intersection ahead, and my heart rate quickly races to cardio speeds. But before I can grab the wheel and slam the brakes, the Nissan smoothly brings itself to a stop, waiting for traffic to clear before it starts up again.
Welcome to the brave new world of autonomous vehicles. Self-driving cars have long been the stuff of science fiction—and of science itself, if you consider some of the little rovers now creeping along on the surface of Mars with only the most occasional instruction from operators at NASA. In recent months, a number of manufacturers have announced plans to bring the technology back down to Earth, something that would realize a fantasy that tantalized even the great Leonardo da Vinci. As much tinkerer as artist, he reportedly designed a robotic cart five centuries ago, though it wasn’t until the New York World’s Fair of 1939 that General Motors captured America’s imagination with the autonomous car concept on display at its Futurama exhibit.
One only has to consider that roughly 35,000 people are killed on U.S. roads each year to understand why some are so excited about the idea of taking the task of driving away from people who might also be texting, talking on a cell phone, grabbing a meal on the go and occasionally even falling asleep behind the wheel. “We have a clear vision of accident-free driving,” says Steffen Linkenbach, director of engineering systems and technology for the German automotive supplier Continental AG. Others suggest that autonomous vehicles could eliminate highway fatalities entirely.
Not everyone is quite so confident. Even proponents admit it will be a daunting task to get to the stage at which we can slip into the rear seat, tell our car where we’re headed and lie back for a quick catnap before reporting for work. It’s one thing to program a prototype autonomous Leaf to navigate around a closed course, admits Maarten Sierhuis, co-director of Nissan’s self-driving vehicle program, but “we need to be able to drive any intersection anywhere and at any time to be fully autonomous,” stresses the former NASA rocket scientist.
Sierhuis and his team are going to have to work fast. Nissan recently promised to put its first fully autonomous vehicle into production in 2020, and during a presentation in California last August—during which I got my first chance to ride, er, drive, er, copilot a self-driving car—the Japanese maker’s global product czar Andy Palmer declared that, “Basically, within a decade, you’ll have this technology available in every car” bearing the Nissan badge.
Over the past decade, DARPA, the Pentagon’s research arm, has toyed with autonomous technology, hoping to someday be able to go to war without putting human soldiers in harm’s way. But there’s a big difference between a tank and your sedan or SUV. Curiously, it wasn’t an automotive manufacturer that ushered in this brave new world. On October 9, 2010, the Silicon Valley tech giant Google made a terse announcement on its corporate blog noting that it had begun testing a self-driving vehicle of its own design. “Our automated cars, manned by trained operators, just drove from our Mountain View campus to our Santa Monica office and on to Hollywood Boulevard. They’ve driven down Lombard Street, crossed the Golden Gate Bridge, navigated the Pacific Coast Highway, and even made it all the way around Lake Tahoe. All in all, our self-driving cars have logged over 140,000 miles. We think this is a first in robotics research.”
Google remains one of the pioneers in the field and is reportedly looking for a manufacturer that might want to partner with it to put its technology into production. But by now, it seems, virtually every automaker—and a number of their suppliers—have begun working on autonomous systems of their own, and they’re getting there, step-by-step. It’s been little more than two decades since the first active safety technology went into production. Unlike traditional, or passive, safety systems, which simply aim to reduce injuries and fatalities when a crash occurs, the concept behind active safety is to prevent an accident in the first place. Indeed, one only has to get into the newly redesigned Mercedes-Benz S-Class to see just how far things have come.
The German flagship sedan offers the latest incarnation of the earliest active safety technologies: antilock brakes, traction control and stability control, all designed to maintain a vehicle’s grip on the road while keeping it pointed in the direction the driver intends. But the big Benz can do far more with what Dieter Zetsche, boss of the Mercedes brand and CEO of parent Daimler AG, calls “sensor fusion.” A fully loaded S-Class is equipped with the sort of monitoring equipment that would bring pride to the NSA, the government’s high-tech spy agency. There are cameras, sonar and radar sensors, even a night vision system, together maintaining a 360-degree vigil focused on the world around the sedan.
There’s the latest version of the Mercedes Distronic technology which lets the new S-Class keep pace with the traffic flow, come to a complete stop in a tie-up and then start rolling again. It will trigger the brakes if a collision seems likely and even turn on the flashers when one does occur. The latest in lane-departure warning systems will even take control if a drowsy driver starts to drift into an adjacent lane. The vehicle will constantly watch for pedestrians and animals and in urban driving can bring the big sedan to a complete halt if one walks in front of it. Out on a highway, the newest-generation Night Vision system will help spot not only pedestrians, but deer and other animals. The system will even flash a spotlight to warn people out of your path. There’s also a vision-based system called Magic Body Control that constantly scans the road surface ahead, preparing for dips and bumps by rapidly adjusting the stiffness of the car’s shock absorbers. The technology in the 2014 S-Class, declares Zetsche, is “paving the way towards (fully) autonomous driving.”
Add accessories like the $2,260 Night Vision Assist Plus and $900 Surround View Camera, and the new Mercedes S550 quickly goes north of $100,000, putting it out of reach of all but the most affluent—and safety-minded—motorists. When you consider that there’ll almost certainly be a raft of additional devices onboard a fully autonomous vehicle, it might seem like the technology will be available to only a select few. The good news is that just as with consumer electronics, microprocessor-based automotive safety technology is rapidly plunging in price. Consider that when it was first introduced in the 1990s, electronic stability control was a $2,000-plus option offered on only a handful of luxury cars. Today, it’s mandated on all new vehicles sold in the U.S., including the $15,500 Kia Rio. CrossTraffic Alert, which spots approaching cars as you back out of a mall parking spot, for example, was an exclusive on the BMW 7-Series just a few years ago and is now becoming commonplace on mainstream midsize and even some compact models. Subaru and Nissan have both launched vision-based systems that handle many semiautonomous tasks on the cheap.
A growing number of new vehicles, such as the Ford Taurus, now feature auto-parking capability. The technology will scan for a spot your vehicle can slip into and then smoothly steer into the space—though most systems still require the driver to use the brake and throttle. But Ford and Toyota recently showed off prototype “valet” parking systems that would allow a motorist to get out of the car, press a button and have the vehicle park entirely on its own.
By the time autonomous technology truly takes off, it’s expected that there’ll be an extensive vehicle information network in place, including both a roadway infrastructure and so-called V2V or vehicle-to-vehicle systems that could, for example, alert your car to upcoming weather or traffic problems. You might also be able to have your vehicle drop you off at work then drive off to a parking spot reserved for it automatically at a nearby “connected” garage. You’d tap a smartphone app to have it come back to pick you up later in the day.
The pace with which all this technology is hitting the road is surprising even proponents of autonomous driving, some of whom didn’t really believe the concept would be ready for prime time before mid-century. “I have to admit that I did not think mass production was doable by 2020, but I am beginning to change my mind,” says Mark Campbell, a Cornell University professor of mechanical and aerospace engineering who leads a research group that studies and builds autonomous systems. “The power of Google in this arena has been tremendous, as Google takes an approach that is more research-driven, cutting edge and riskier than the car companies.”
Google may have given autonomous driving its first big push, but the number of manufacturers signing on is increasing rapidly. Along with Nissan, Tesla and General Motors are among those who have announced plans to get into production by the end of the decade or soon afterwards. And others are expected to follow.
Reducing the annual carnage on our highways is just one of the reasons advocates are so excited about self-driving vehicles. Another key goal is to improve the flow of traffic, especially in cities like New York, Los Angeles, Atlanta and Washington, D.C., where most highways already are well past the maximum volumes they were designed for. In fact, it is promised, autonomous vehicles could save tens of billions of dollars—and reduce the impact on the environment—by curbing the need to constantly add new lanes and more roads. In turn, by reducing stop-and-go driving and jack-rabbit launches when the light turns green, self-driving cars could deliver substantially better fuel economy.
So, what’s not to like? Quite a lot, according to some recent public opinion polls. A survey conducted for the website CarInsurance.com found 20 percent of American motorists would rather give up driving entirely than be forced to depend upon an autonomous vehicle. And a full 75 percent said they wouldn’t trust the technology to take their children to school. Two out of three respondents said they didn’t believe computers could make driving decisions as effectively as a human—something you might question after spending an hour or two commuting every day. That said, the number of motorists who now like the idea of relinquishing driving duties is beginning to increase, according to the site’s managing editor Des Toups. And, “Our survey shows cheaper insurance will greatly influence consumer acceptance.”
Fewer crashes should mean substantially lower rates. At least, that’s another one of the promises made by autonomous-driving proponents. But will your insurance company really give you a discount? Indeed, will Progressive, Allstate and State Farm even cover self-driving cars? That’s certainly a big concern for Nissan, product chief Palmer warning that, “We may not see autonomous vehicles in the U.S., at least not initially” if that would mean an endless stream of lawsuits every time something happens involving a self-driving car, no matter who was really at fault.
As a result, we could see the technology reach the road in Asia first. China could certainly use it considering that fast-growing automotive market officially recorded 60,000 traffic fatalities last year—though several studies contend the government’s numbers are grossly deflated. The World Health Organization puts the likely tally at around 235,000 annually.
Eventually, autonomous driving seems destined to reach U.S. roads. But the payoff could be slower than advocates would like you to believe. Remember, there are currently well over 200 million vehicles on U.S. highways, the average one now about 11 years old, according to industry data. Even if federal regulators could require all new vehicles to adopt autonomous driving technology, say, starting in 2025, it would likely be mid-century before the vast majority of the fleet would be replaced. So, for years to come, your autonomous vehicle would have to cope with folks still texting, eating and doing all sorts of things that take their eyes—and minds—off the road.
The reality, the experts predict, is that autonomous vehicles will phase in even more slowly. And, for at least the early years, you won’t be able to curl up for that nap. Nevada, the first state to officially create a special class of licensing for self-driving cars, currently requires that an “operator” remain behind the steering wheel ready to take over immediately in an emergency. That’s a troubling issue that researchers have yet to resolve. Just how well—and how fast—could a human driver properly regain control if, for example, there were to be a massive failure of a vehicle’s autonomous control system? Some data suggest it would take about seven seconds. But that’s assuming that the person behind the wheel isn’t texting or eating, reading or struggling to stay awake and focused after a few hours of having nothing to do while riding copilot on a 300-mile journey.
So, there are those who question how fast autonomous technology will take over—if it will become the norm, at all. Rupert Stadler, the CEO of Volkswagen’s luxury arm Audi, expects a very gradual roll out, cautioning the implementation of autonomous driving will be “much more…an evolutionary process than a revolutionary process.” Jim Lentz, the top U.S. executive at Toyota, is perhaps even more cautious. “We view autonomous cars a little differently than some others,” he explains, adding that, “We really see it as a copilot type car, not as a self-driving car,” one that takes the semiautonomous technologies now becoming common a few steps further but which still leave the human in charge. But that, Lentz recently told the Associated Press, could still have a significant, positive impact. “As we look at boomers and they start to retire, the ability to have cars that can enhance their capabilities, it’s going to allow them to drive much longer.”
Despite the headlines and all the promises, there are still plenty of speed bumps ahead in the race to make autonomous driving a reality. Few would be surprised to see Nissan and its competitors push back their end-of-decade targets. But that said, the pace at which semiautonomous technologies are coming to market is rapidly accelerating, so even if you won’t be turning driving duties over to your car in the near future your vehicle will increasingly be second-guessing your every decision—and the result is likely to be much safer highways.
Paul A. Eisenstein is publisher of the website The DetroitBureau.com.