The Google self-driving car program passed a milestone on Wednesday as it marked more than two million miles of driving on public roads. Google’s Alphabet Inc. division started working on autonomous cars back in 2009 and now has over 60 self-driving cars on the road.
Currently, they’re clocking around 25,000 miles of autonomous test driving every week putting traditional automakers to shame. But it’s still unclear exactly Google plans to do with all of the data it has collected over those two million miles.
They haven’t said whether they plan to have only their own car, or if they plan to use this research to partner with traditional automakers or even new players to get autonomous cars to the public. What they have said is that they believe fully autonomous cars with no need for a driver, steering wheel, or pedals, will be ready by 2020.
Traditional automakers are taking a different approach. Rather than holding back until they’re ready to release a fully autonomous car, they’re introducing bits and pieces of autonomous technology as it’s available. You see it in features like emergency braking, lane keeping assist, and adaptive cruise control. These are all pieces that need to be mastered before a fully autonomous car stands a chance.
For now, Google is using their cars as giant data collectors. There are a vast number of variables to consider even when driving the same route. They’re trying to understand those variables whether it’s the difference between a person, a dog, and a bicycle or more complex social cues.
It’s those social interactions that are the toughest to master. Think about the kind driver who waves to let you into heavy traffic. There’s no law that covers doing this, but it’s something everyone understands. It might come with a blink of the headlights, a wave out the window, or just a nod from the driver to let you know he’ll wait and let you go. How does a car understand those kinds of cues?
This makes the number of miles less important than the type of miles. City traffic poses far more opportunity for social interaction where highway miles are more uniform and less complex.
Google will continue to rack up those miles, collecting data as it learns from every interaction. The faster it learns and the more data it collects, the sooner fully autonomous cars will find their place in the automotive landscape.