I put this in NQS because it isn't strictly a scientific subject. It is a technology subject. There is a lot of science behind self driving cars, but the science has already been worked out and is being applied.
Wired Magazine has an article about self driving cars in this months issue (20.02). There is some discussion about the different approaches to developing the cars. They point out that Google, which has a really impressive car, is looking at it from a data view point, while most of the major car companies are looking at incremental improvements to the car the way it is. In any case they are beginning to actually work out some pretty good ways of having a car drive itself. They also still have a ways to go before they really have a car that can be just told to go to such and such a place and then let it loose.
One thing they talk about is that for most people, except for advertisers, the driving isn't what cars are all about. They are basically to get some place and the driving is just an interruption. If we didn't have to drive we could spend more time working or carrying on our networked social life.
One point about the new breed of self drive cars is that they don't need special roadways. In the past we have heard a lot about what it would take to make self-drive cars work, and that usually included reworking the roads with special structures to guide the cars. These cars are fully autonomous, they work with roads the way they are.
They also point out that one impediment is going to be liability issues. If a self drive car is involved in an accident who is responsible? That can open up a whole new gold mine for liability lawyers. (They didn't say it that way, I did). So insurance companies are a bit leery of fully automatic cars.
In the mean time the man from GM is predicting autodrive cars by 2020. I suspect that is a bit optimistic, but it certainly looks as if they are coming.
Bill Gill
