A map of Google's self-driving car accidents

Last week, when the California’s DMV released a year’s worth of reports for accidents that involved self-driving cars, it turned out that all of the accidents were caused by humans, as opposed to software glitches on the cars’ parts. We wondered what else self-driving car accidents might have in common so we decided to map accidents involving Google’s cars, based on information released by the DMV as well as Google itself, which released a report this year on all of the accidents it’s had so far.

Google has had 16 accidents total since it started putting its self-driving Lexuses and Priuses on the road in 2009. The first accident, according to Google’s self-reports, occurred in May 2010. Google’s cars have primarily stuck to roads around the Bay Area (though they recently began testing in Austin, Texas as well). It turns out that the majority of its accidents have occurred “near home,” within 20 miles of Google headquarters. This map of 11 accidents, each represented by a blue dot, shows a cluster of mishaps in the heavily trafficked commuter roads near Google’s campus in Mountain View.

That most of the accidents happened in downtown Mountain View makes sense. Urban driving is a much more difficult problem to solve because of the unpredictability of pedestrians, traffic lights, cyclists, and stray animals. Highway driving is monotonous by comparison, and much easier to program. To date, Google has not reported an accident in Austin, Texas, according to the company.

To give you a better perspective on where you’re most likely to run into a self-driving car, perhaps literally, here’s that neighborhood on a larger map of the Bay Area. On this map, we’ve mapped the other five accidents Google has had, which all had more general address information included in their accident reports.

In one case, we actually have real-time documentation of an accident. In July, one of Google’s Lexus autonomous vehicles was driving itself toward an intersection. There was a green light, and cars in front of the Google car started breaking. The car behind Google’s didn’t start breaking in time and slammed into it at 17 mph, according to a recent blogpost detailing the event. The accident was recorded by the car’s 360-degree “eyes” and you can see it for yourself in the video below:

It’s obvious the human driving behind Google’s Lexus was distracted. They seemingly had plenty of time to brake, but they didn’t.

“Drivers have hit us 14 times since the start of our project in 2009, including 11 rear-enders,” said Google’s Chris Urmson in a July 2015 blog post. In a 2011 collision, a human Googler was to blame while operating a driverless car in manual mode. In August 2015, after that blog post, a Google car got hit yet again by a human-driven car

It’s hard to blame the poor humans. Google’s cars, after all, get a lot of more practice driving. Each week, they log about 10,000 miles, compared to American humans who drive about 15,000 miles per year. Given that 9 people die and more than 1,100 people are injured every day due to car crashes involving distracted drivers, according to the Centers for Disease Control and Prevention,  Fusion’s Kevin Roose recently argued that we should ban human driving and let the robots take over.

Daniela Hernandez is a senior writer at Fusion. She likes science, robots, pugs, and coffee.

 
Join the discussion...