Google auto project is about making the perfect driver, not a car

Despite building and operating a fleet of self-driving vehicles, Google says its end goal is not to make autonomous cars, but to create autonomous drivers to pilot them.

Google and parent company Alphabet have been working on autonomous car technology since 2009 and in the following seven years millions of miles have been covered with computers at the wheel.

First the company used modified Toyota and Lexus cars, before building its own pod-like vehicles, designed to operate entirely without human control. Now, it is preparing to work on a fleet of 100 Pacifica minivans built by Fiat Chrysler.

Google's self-driving software is already advanced enough to control a vehicle, understand road layouts, follow, avoid and give way to other motorists when needed, and even understand the hand signals of cyclists Now, Google wants to design its suite of sensors and algorithms to work in any vehicle, like how humans quickly adjust to unfamiliar hire cars.

Speaking to Recode, Google's head of self-driving technology, Dmitri Dolgov, said: "It doesn't matter [what car is being driven]. We're building a driver. We've been on Prius, Lexus; we have our own prototype, and we're now working with Fiat Chryslers on a new platform.

"As far as the software is concerned, it's the same thing. It's like you getting into another car. You get a rental – maybe it's a little bit longer, and it doesn't quite handle the same as your own car – it takes you time to get used to, but the core tasks transfer."

As well as using a mapping database, lasers, cameras and radar to drive safely alongside other traffic, Google's cars also share what they learn with each other via the cloud. This helps them to learn much more quickly than humans and be better prepared for scenarios they are yet to encounter. Despite growing fears of internet connected cars' vulnerability to hacking, Dolgov says there is little a hacker could do if they gained access to Google's cloud network.

"We limit the amount of information that can be shared, and how that affects other cars. If you have complete control of one vehicle, maybe you can create a construction zone [and thus warn the car to look out for temporary road changes] that doesn't actually exist in the world, but that's the extent of it. There's no way to control the local driving behaviour."