Google’s self-driving cars can tackle the roads in Mountain View, California so well because the company has created a virtual track of the city’s streets
Google then loads that route data into the car’s memory before it drives off. That way, the car knows exactly what turns, stop signs, curbs, and other obstacles it needs to expect.
“Rather than having to figure out what the world looks like and what it means from scratch every time we turn on the software, we tell it what the world is expected to look like when it is empty,” Andrew Chatham, Google’s self-driving car team’s mapping head told The Atlantic. “And then the job of the software is to figure out how the world is different from that expectation. This makes the problem a lot simpler.”
That means Google has a long road ahead of it in making self-driving cars work nationwide, yet alone worldwide. But the company isn’t worried.
“It is work,” Google self-driving car lead Chris Urmson told The Atlantic, “but it is not intimidating work.”
Since launching the driverless car program in 2009, Google’s cars have logged over 700,000 miles. It also recently set out to tackle autonomous driving on city streets.
On Tuesday, Google revealed a prototype driverless car without pedals, breaks, or a steering wheel. Google plans to deploy at least 100 of these this year.
Check out the self-driving car software in action.