Google's self-driving cars have driven over 2 million miles — but they still need work in one key area

Rated 4 out of 5 stars based on 4 votes

Google has been developing its cars since 2009 and one of its favorite stats to share about the project is that its self-driving cars have driven over 2 million miles. Google, which spun out its self-driving car unit into an independent company called Waymo last week, wrote on the Waymo website that the cars now have "the equivalent of over 300 years of human driving experience, largely on city steets."

That kind of mileage shouldn't be taken lightly — Google's cars are extraordinarily perceptive and can recognize objects that can be difficult for self-driving cars to see, like bicycles.

But at a time where Google is feeling growing pressure from competitors like Uber and Tesla, the tech giant has yet to test its self-driving cars in cold weather or snowy conditions. As anyone hailing from the East Coast or Midwest can attest, driving in snow is a required skill.

Snow poses a particular set of challenges for self-driving cars because it can confuse the systems they rely on to get around, like cameras and lidar, a sensor that uses lasers to map the car's surroundings so it can "see" the world.

When there's snow on the ground, cameras and lidar have a difficult time seeing lane markers, which cars rely on to prevent lane drift and navigate safely. Snow can also make it more difficult to detect unexpected obstacles.

"Heavy snow and rain tend to confuse lidar sensors and also cameras," John Dolan, principle systems scientist at Carnegie Mellon's Robotics Institute, previously told Business Insider. "So you end up having some problems."

Automakers and startups investing in self-driving cars have generally been forthcoming about the challenges posed by snow fall and how they are planning to address them.