The Technology Behind Google's Autonomous Cars

Over the weekend, Google has unveiled one of its most ambitious projects to date, autonomous vehicles. The technology company has a fleet of seven cars which have clocked in 140,000 miles on public roads in California as well as in more controlled environments.

"Our automated cars use video cameras, radar sensors and a laser range finder to 'see' other traffic, as well as detailed maps (which we collect using manually driven vehicles) to navigate the road ahead," Sebastian Thrun, the Google engineer who leads the project, explained.

"This is all made possible by Google’s data centers, which can process the enormous amounts of information gathered by our cars when mapping their terrain," he added.

The automated cars use three types of detection sensors to make sense of the world around them.

One technology used is LIDAR, or Light Detection And Ranging, a type of radar that uses lasers instead of radio waves for detection. LIDAR is incredibly precise, compared to regular radar technology, but is only effective at short and medium distances, up to 80 meters or so.

The LIDAR sensor is mounted on top of the cars, similar to the camera equipment used by Street View cars.

The cars, six hybrid Toyota Prius and one Audi TT, also have four regular automotive radars which have a greater range than LIDAR. Three radars are mounted in front and one in the back.

They also use a high-definition video camera to detect motion, like pedestrians or cyclists, and 'read' things like stop signs and traffic lights. This camera is located near the rear-view mirror.

To complete the range of sensory data gathered, the cars also come equipped with GPS devices and accelerometer sensors which register motion.

All of this links up to Google's data centers which already have a vast amount of mapping data. The data comes from a variety of sources, but a lot of it comes from the cars themselves which map the terrain they drive through and send huge amounts of information back to the Google data centers.

For anyone following developments in the rather narrow field of autonomous driving, this is nothing new, most projects to date relied pretty much on the same equipment. The only real difference is the cloud, previous projects have relied on on-board hardware for all their processing power.

However, the advances that Google has made are in the algorithms and, of course, the real world testing that these cars have done. It is by far the biggest such experiment to date and the amount of data gathered comes in very handy in improving the technology.

Hot right now  ·  Latest news