The tech behind Google's driverless car
Google's self-driving car got its license this week as the state of Nevada became the first in the nation to license the company's vehicles.
And while a computer-driven car may seem unsettling, the technology represents a potential leap forward in auto safety.
More than 30,000 people are killed each year in crashes despite huge advances in auto safety. The overwhelming majority of those crashes are caused by human-driver error.
Computer driven cars could reduce traffic deaths by a very significant degree, said David Champion, head of auto testing at Consumer Reports, but only if all is under control of a driver.
That's because humans are better at predicting the behavior of other humans than computers could ever be, he said.
For the foreseeable future, human "drivers" will continue to bear the ultimate responsibility even in Google's self-driving cars. This means you won't be able to lounge in the back seat and check email on your way to work. You'll still have to sit in the driver's seat and pay attention.
Self-driving cars, like Google's, use sensors to watch cars, pedestrians and other obstacles. They combine a number of technologies that are already available on cars today -- including GPS tracking, wheel motion sensors and radar -- with additional technology and sophisticated software that allow the car to read street signs and signals and actually drive itself through traffic.
Google's cars, modified Toyota Priuses, are still in the testing stages and aren't available to the public. But some so-called "driver assistance" technologies are already helping to lower traffic deaths in cars you can buy now.
Electronic Stability Control which uses computers to help drivers maintain control during abrupt maneuvers, has been shown to reduce fatal crashes by as much as a third.
ESC is now required on all new cars but was first used, on a wide scale, on SUVs. That's why, last year, statistics showed top-heavy SUVs to be less prone to roll over in real-world crashes than regular cars.
Beyond that, there are various other "driver assistance" technologies.
Blind spot alerts warn drivers of cars in adjacent lanes and forward collision alerts sound an alarm when a driver is closing in too quickly on a car ahead.
GM's new Cadillac XTS, for instance, will brake automatically if a driver fails to respond to an imminent collision. Nissan's Infiniti division has a several models that provide slight braking to nudge a vehicle back into its lane if it begins to drift out.
Many luxury cars are now also available with "active cruise control" that allows a car driving at highway cruising speeds to automatically maintain a safe following distance behind the car ahead.
With these systems, drivers set the active cruise control to a certain speed. If there's a slower car ahead, the cruise control will automatically slow the vehicle down to maintain a safe distance between the two cars. Once the slower car moves away, active cruise control will accelerate to the higher preset speed. This acceleration can be startling to drivers unfamiliar with the system.
There is at least some evidence, however, that "driver assistance technologies" do work. A recent study by the Highway Loss Data Institute, an insurance industry group, indicated that the forward collision avoidance system in the Volvo XC60 helped reduced accident claims by 27%. Volvo's system warns the driver of an impending collision and applies the brakes if the driver takes no action.
0:00 /2:05See the disappearing garage in action
One technology the Google car doesn't utilize, but which would help make self-driving cars much more effective, Champion said, is vehicle-to-vehicle communication. So called V2V communication uses transmitters to send and receive signals that tell other cars where each car is, where it's headed and how fast it's moving. The devices can also communicate with transmitters along the road.
V2V is already in advanced stages of development by a consortium of automakers and the federal government's National Highway Traffic Safety Administration.