The third layer uses data from human driving behaviour to predict what the car’s next movements should be, or, to make it act in a more “human” way. For example, just because the speed limit on a road is 60mph, it doesn’t necessarily mean the car should be driving at 60mph.
The maps have got all their information from vehicles driving all over the globe, assimilating street level, panoramic and 360-degree imaging. IN the UK alone they cover 65,000 miles a year.
At the top of the vehicles sits a whopping 5ft rig. A small cylinder, like a fizzy-drinks can, spins rapidly as the car is travelling along. It’s the LiDar sensor itself, capturing 700,000 3D data sections a second. It captures detail like the edge of a curb or height of a tree and works within a margin of error that’s 10cm (which is quite amazing when you consider how much data there is to capture, but slightly worrying in that any margin of error when driving is unwelcome).