This week, one of Google’s autonomous vehicles was involved in a minor fender-bender when the car backed into the side of a passing bus. While the self-driving cars have been involved in traffic accidents before, the company has always maintained that the crashes were caused by human drivers sharing the road with the vehicles. In this case, Google is accepting “partial responsibility” for the accident – a first for the company – and is working on changing its software to avoid such incidents in the future.

Continue reading below
Our Featured Videos
google, google x, x lab, self driving car, google car, autonomous vehicle, driverless car, mountain view, california, car accident, crash

While it sounds worrying, the accident wasn’t serious. A self-driving Lexus traveling eastbound on El Camino Real in Mountain View, CA encountered sandbags positioned around a storm drain at a stop light. In order to go through the intersection, the car had to back up slightly to maneuver around the sand bags. Behind the vehicle was an approaching city bus, which both the car’s AI and its human test driver anticipated would slow or stop to make way for the car. Instead, the bus continued to travel at 15 mph, and the automatic vehicle grazed the side of the bus at about 2 mph.

Related: Google’s autonomous cars will drive themselves around Mountain View this summer

There were no injuries during the accident, although the Google car sustained some damage to its front fender, left front wheel, and one of its driver’s side sensors. While Google has admitted the car (and test driver) made the wrong call in this situation, the company is quick to emphasize that “This type of misunderstanding happens between human drivers on the road every day.” While that’s undoubtedly true, it may not be reassuring to those who find the idea of a self-driving car unsettling.

Still, Google is already taking measures to help avoid a repeat of the incident in the future. Its cars have been reprogrammed to be more cautious around large vehicles like buses, and not assume that they will yield. This is part of an ongoing effort to train its cars to drive a bit more like a human would to avoid dangers on the road.

+ Google Self-Driving Car Project

Via The Verge