Google's autonomous Lexus self-drives into the side of a bus
Google's perfect self-driving record has received its first little blemish, with one of the company's autonomous vehicles colliding at low speed with the side of a bus in Mountain View, California.
In the six years that Google's self-driving cars have roamed city streets testing out its autonomous vehicle technology, they have been involved in at least 17 accidents. But the company has never conceded that its vehicle was the cause of a collision.
In an incident in mid-February, a Google Lexus autonomous SUV was making a right-hand turn when it encountered a set of sandbags blocking its path. After coming to a stop and waiting for a break in traffic, it attempted to veer back into the center of the lane and hit the side of a bus at around 2 mph (3.2 km/h). The bus was traveling at 15 mph (24.1 km/h).
According to accident report filed with the California Department of Motor Vehicles (DMV), "the Google AV test driver saw the bus approaching in the left side mirror but believed the bus would stop or slow to allow the Google AV to continue." Nobody was injured in the incident but there was some damage to the car's left front fender, left front wheel and one of the sensors. A witness video uploaded to YouTube can be viewed here.
In what marks the first admission of fault from Google with regard to its autonomous driving tech, according to The Verge its forthcoming monthly report states, "in this case, we clearly bear some responsibility, because if our car hadn't moved there wouldn't have been a collision. That said, our test driver believed the bus was going to slow or stop to allow us to merge into the traffic, and that there would be sufficient space to do that."
Given that Google's self-driving cars have logged over one million miles (1,600,000 km) of autonomous travel since the program began in 2009, this is one pretty minor speed hump on one very long road. The company says it has reviewed the recent incident along with thousands of variations and made refinements to its software, which will allow it to better understand that large vehicles are less likely to yield than others.