Newly released video shows recorded by cameras on the bus show the moment a Google self-driving car learned the hard way not to tussle with a public bus.
Sourced through Scoop.it from: www.dailymail.co.uk
Google is doing the right thing in being open about the training of driverless cars. What this situation tells me though is that as the ‘drover’ Priscilla tells us in the video, there is still a long way to go.
Where situations are cut and dried and objects or ‘boxes’ that could impede the journey are easily identifiable, the Google car at speeds of 15 – 40 mph does pretty well and I would expect that.
However it’s the wild cards that worry me. As anyone in the traffic industry knows, no two events are the same, there are loads of variables and the most dangerous variable is the random human. The example of a cyclist who repeatedly signals a lane change, changes their mind and does it again, is great. But most cyclists I see changing lanes don’t indicate at all.
The crash happened because of sandbags on the side if the road. So what other random things happen? How about cyclists who decide to ride illegally 3 abreast on a windy road just as the car behind them comes to a blind corner? What about the situation last week where a woman caused a traffic jam in Texas by climbing onto the cab of a truck and started dancing naked.? https://www.rt.com/viral/334884-naked-woman-dancing-truck/
The fact is that there are so many random events and combinations of events that it is difficult for humans with years of experience to know in a split second how to handle them.
I have no doubt that in some shape or form driverless cars will be a common form of transport. For the foreseeable future that may be under very controlled circumstances. It’s all very well to say that driverless cars have been involved in very few accidents. But what is the ratio of driverless cars to human controlled cars?
I do credit Google for being open through this process, that is how to win trust for a technology that improve our lives….eventually.