Google uses its giant data centers to simulate 3 million miles of autonomous driving per day, the company has revealed in its monthly autonomous driving status report. That's a really long way — like driving more than 500 round trips between NYC and LA — but it actually makes a lot of sense. Americans drove some 2.7 trillion miles in the year 2000 alone and Google needs all the data it can get to teach its cars how to drive safely.
The real advantage comes when Google's engineers want to tweak the algorithms that control its autonomous cars. Before it rolls out any code changes to its production cars (22 Lexus SUVs and 33 of its prototype cars, split between fleets in Mountain View and Austin), it "re-drives" its entire driving history of more than 2 million miles to make sure everything works as expected. Then, once it finally goes live, Google continues to test its code with 10-15,000 miles of autonomous driving each week.
The simulations also allow Google to create new scenarios based on real-world situations — adjusting the speeds of cars at a highway merge to check performance, for example. Engineers can then design fixes and improvements and check them in the simulator, ensuring that things are as operating as safely as possible before Google's cars make it out onto real roads.