The quest for perfecting the science behind self-driving hits a new high with Google reporting a mileage record of two million miles for their computer-mediated vehicles.
This was recently confirmed earlier by Dmitri Dolgov, head of Google's self-driving tests. The feat is a result of at least seven years of repeatedly experimenting with the idea of a software-led mechanism for driving.
"In 2009 we were just this crazy research project working on a science fiction idea. And look at where we are today," Dolgov said during an interview with .
Of the two million miles covered by their self-driving Lexus cars, at least 90% of the areas covered by the company's test drives include city streets. When converted to actual time, the total work covered by Google equates to at least 300 years of human driving experience.
This is a huge achievement for the company, who has remained on the down low while other companies such as Uber and Ford announcing and rolling out autonomous driving services to the public.
However, Dolgov believes that this is still far from actually achieving the main goal of the technology.
In an interview with , he thinks that there are still a lot of work to be done as he believes that as of now, autonomous driving remains to be hugely ambiguous.
"I find the terminology is over-loaded, and there are big gaps between what's out there and what's working. But it's exciting," Dolgov said.
Currently, the Google team is still in the middle of calibrating digital data that they gather from their equipment so that it can be accurately translated into realistic interpretations, despite how random it is. This is opposed to competitors who are reliant on developing sophisticated algorithms and simply relying on those for autonomous driving technology.
"If there's a kid chasing a ball into the street, the car has to understand that's different from a jogger in the road. Reacting to the world is the core of the challenge, reacting smoothly," he adds.