How these robotic vehicles took off.
After attending the 1964 World’s Fair, the science-fiction author Isaac Asimov wrote an essay in The New York Times imagining a visit to the World’s Fair 50 years in the future, in 2014. Among his predictions: “Much effort will be put into the designing of vehicles with ‘robot-brains’—vehicles that can be set for particular destinations and that will then proceed there without interference by the slow reflexes of a human driver.”
Asimov got some of the details wrong (he thought the cars would ride suspended on compressed air), but most of his prediction proved accurate: much effort is, indeed, now being put into the design of robot cars, thanks largely to Google. Earlier this year, the company revealed a prototype of a fully driverless car, an adorable machine without a steering wheel or pedals that tooled around its campus in Mountain View, California.
Google’s achievement draws on the ideas of computer scientists, roboticists, and automotive engineers who have been working on autonomous vehicles for decades. And the goal is not just to realize our science-fiction dreams: driverless cars might alleviate congestion, ease demand for parking, and reduce crashes, one of the leading causes of death in the United States.
Early efforts were not really robot cars at all, but highway-automation systems. Back in 1956, for instance, GM introduced a Firebird II concept car that would be guided by a hypothetical electric highway of the future.
In the individualist 1980s, though, the cars took control. As autonomous vehicles like Stephen King’s Christine and Knight Rider’s KITT graced the big and small screens, researchers’ efforts began to bear fruit. A team at Bundeswehr University Munich transformed a Mercedes van into a self-driving vehicle called VaMoRs, and the Carnegie Mellon Robotics Institute turned a Chevrolet panel van into the first in its line of Navlab robot cars. (Why vans? To store all the computing equipment necessary to operate the vehicles.)
In 2004, the Defense Advanced Research Projects Agency launched its Grand Challenge series, a multimillion-dollar competition for autonomous vehicles, giving roboticists scattered across American universities and companies a chance to go toe-to-toe. A rivalry developed between Stanford and Carnegie Mellon, which traded the top spots in the first two years a prize was awarded.
The early Grand Challenges resembled the Fast and the Furious films more than government research experiments—and Google took note of the showmanship. Sebastian Thrun, the leader of Stanford’s winning team, who took a leave from the university in 2007 to work on Google Street View, later founded the company’s self-driving-car project.
Rather than setting researchers up to compete for grants, space, funding, and all the other quotidian trials of university research, Google just hired many of the best from Stanford, Carnegie Mellon, and elsewhere, and gave them access to the company’s massive array of computational power and collected data.
Google’s autonomous vehicle is still years away from widespread use: it faces both technical and regulatory hurdles. In September, for example, California began enforcing new rules requiring autonomous vehicles to allow drivers to take control in an emergency, and Google has said it will modify its test cars to comply with the regulations.
Still, the Google car offers lessons about how science fiction can become fact. In a sense, Google’s self-driving car is more of a parts-assembly project than it is a dramatic new vision for human transport. The company’s real breakthrough was bringing together researchers and the existing technologies that underlie their effort—computer vision, digital mapping, and more. Perhaps this is what Asimov overlooked in his vision of 2014: the future is less about technologies themselves than it is about the organizations with the means and the will to put them into practice.