It’s a late summer day, and I’m sitting in the driver’s seat of a BMW 3 Series at the Mazda Raceway Laguna Seca in Salinas, Calif. Sitting, not driving. When I lift my hands from the wheel at the beginning of this 2.2-mile course, the car accelerates to 75 mph almost instantly, pushing me and my passengers—BMW engineers and executives—into our leather seats. The car’s computer brain, using satellite signals to navigate the track, is in control.
“Wait until you see what’s coming up,” says Tom Kowaleski, a BMW spokesman, as we head for the Corkscrew, a steep, tight S-curve and the scene of numerous YouTube crash videos. We hit it at about 40 mph, and I have to sit on my hands to keep them from grabbing the wheel back from the machine. The executives chuckle.
This 3 Series is part of BMW’s ongoing efforts to improve the technology behind driverless vehicles and understand how computerized chauffeurs might be used in the real world. Similar projects are under way at General Motors, Volkswagen, Google, and at research labs around the world. While the current technology is good enough to navigate roadways and recognize obstacles, it needs some refinement before it’s street-safe, says Thilo Koslowski, an industry analyst with researcher Gartner. The component costs also need to come down, he says. Still, there’s enough activity that governments are beginning to think about how to regulate the new smart vehicles. “In 10 years you will see the first kind of autonomous vehicles” on regular streets, says Koslowski. “The privilege of driving is going to be redefined.”
The idea of self-driving cars is almost as old as the car itself. GM’s vision for the future of transportation at the 1939 World’s Fair in New York included driverless cars. Automakers say recent advances in computing power and networking technologies make it feasible to build real ones. Although the experiments vary in their details, autonomous cars generally use GPS to recognize where they are on the road. Cameras, lasers, and radar help them keep their distance from other cars and recognize objects like pedestrians. Superfast processors weave all the inputs together, allowing cars to react quickly.
Proponents say the promise is enormous: Turning the wheel over to computers could lead to less traffic, fewer collisions, and more transportation options for aging societies. The world’s population is predicted to grow 33 percent, to 9.3 billion, by 2050. If that population were to live like Americans, there would be 7.7 billion cars on the roads—up from 850 million today. That enormous fleet would consume 375 million barrels of oil per day, more than five times the global production in 2008, according to John Sterman, a professor at the Massachusetts Institute of Technology’s Sloan School of Management. The traffic problem that already cripples many of the world’s megacities would get a whole lot worse. Bringing the reliability of silicon to the roads would help solve these problems: Autonomous cars could potentially drive at high speeds and close together without fear of wrecks or jams, cutting down on wasted time and gas.
Some auto executives worry about what will happen to their businesses if traffic continues to get worse and driving becomes more trouble than it’s worth. “The freedom of mobility that my great-grandfather brought to people is now being threatened,” said Ford Motor Executive Chairman Bill Ford at a speech earlier this year at the TED technology conference. “Global gridlock is going to stifle economic growth and our ability to deliver food and health care, particularly to people that live in city centers, and our quality of life is going to be severely compromised.”
To avoid that future, car companies from Toyota Motor to Tata Motors have to start thinking of themselves as being in the “transportation services” business, says MIT’s Sterman. “The ones that say we’re in the car manufacturing and sales business are not going to survive the big disruptive transition that is clearly coming.” Automakers are already adding features that cede some of a driver’s control to computers in the name of safety or convenience. Features available in some luxury cars, such as cruise control that uses sensors to maintain a safe distance behind other vehicles or automatic parallel parking, are indicators of an industry starting to think about how we drive, not just what we drive.
Building and servicing intelligent transportation systems, such as the automatic parking feature and real-time traffic feeds, is already a $48 billion industry in the U.S., according to the Intelligent Transportation Society of America. Companies such as Intel and AT&T are eager to grab a piece of that by supplying the microchips and cellular services that help make cars smarter. “Autonomous driving is almost like an evolution of technology rather than a revolution, if you think about the technologies that have been put on vehicles” in recent years, says Chris Borroni-Bird, GM’s director of advanced technology vehicle concepts. His team is working to test the automaker’s EN-V vehicles—autonomous, electric concept cars that made a splash at the Shanghai World Expo in 2010—in the Chinese city of Tianjin.
A major challenge for driverless roadways is finding a way for vehicles to safely and reliably communicate with one another. That requires getting all the automakers and regulatory agencies to agree on a standard. The National Highway Traffic Safety Administration has begun studying various technologies for vehicle-to-vehicle communication and plans to make a decision by 2013 about whether to continue to study the issue and make rules to regulate it. The NHTSA says intervehicle communication could reduce up to 80 percent of vehicle crashes involving nonimpaired drivers. “That really is the moon shot,” says NHTSA Administrator David Strickland. Car-to-car communication could also allow cars to pass warnings about collisions or traffic to each other. Nevada received national attention this summer when the state enacted a new law requiring the creation of regulations governing autonomous vehicles on highways.
State regulations, however, are far from my mind as the BMW I’m “driving” races through the Corkscrew. The car hugs the curve while taking a right turn, and the wheels thump off the track. An engineer reaches over to grab the wheel to keep us on the road. “Sorry about that,” he says, and later tells me the GPS signal was acting glitchy that day. Obviously, there are still issues to work out before we let go of the wheel for good.
The bottom line: Intelligent transportation systems like self-parking cars are already a $48 billion business and a harbinger of autonomous driving.
Higgins is a reporter for Bloomberg News.
325 comments