This is one of my favorite Ted Talks that shows what one of Google’s Self Driving Cars “see”:http://ted.com/talks/view/id/2291
It gets really interesting for me around the 7:50 mark where I came to realize how incredibly complicated and messy the real world is through the lens of a self driving car.
The team at Google essentially has to “teach” the car to make safe decisions instantaneously, as described in the graphic from Google below:
The fascinating part for me is not the car knowing where it is (thats relatively easy) but rather having the car parse and understand the mountain of data in real time from its sensors and make predictive models about how objects will behave. For example, flashing police lights mean the car needs to quickly pull to the curb or a bicyclist with their arm outstretched is likely to slow to turn.
When the car encounters a new situation – or an object behaves unexpectedly – that learning can be immediately processed and shared with the rest of its peers, making the collective wisdom of the network exponentially greater. We’ve already seen some of that innovation in the consumer space with apps such as Waze aggregating traffic speeds from thousands of users and routing traffic in real time through the most efficient routes. Contrast that with (analog) human drivers; when a new rule, regulation, or change comes into effect adoption is never immediate — and in some cases can have devastating drawbacks.
Finally, I think of the unlimited possibilities when it comes to opening transportation options to everyone. Personally, I’ve seen first hand the limitations on personal independence when a senior loses their drivers license due to age or for individuals with visual or cognitive impairments. Can’t wait to see what the future holds.