I agree. I think we will see highways with dedicated 'autonomous' lanes may be, and city routes carved out the same IF we are to get there as soon as people would like / think. Replicating the human driving experience in an AI will take much longer if we chose not to provide a controlled environment for it.
I think that in some areas having hardware built into the road will enable autonomous driving in places where it probably can't happen using video, but in some places where there already are good road markings and places that do not have very bad weather regularly, I think we will be able to see it sooner, though, as I said earlier, possibly restricted to certain places and circumstances. I also totally agree with your comment about replicating the human driving experience taking longer without a dedicated environment... sadly, I think that we as a society are very unlikely to provide a controlled environment (outside of test tracks and very limited places) for this until the autonomous driving was pretty much proven already. Look at how hard it is to get money to repair existing roads or repaint existing highways... the thought that we would block them off and install special devices just for autonomous cars to use seems about as unlikely as you could get right now. it's the classic chicken/egg issue. And the same one that plagued electric cars in the beginning. (who would buy one if there were no public chargers around? we got over that, but it took years!)
However, I don't think it is actually that far away. (Note, I am mostly talking about something like a Level 4 autonomous system here, one that can handle MOST situations on its own, but will ask for help if needed). I think that from a technology POV we are already there in a lot of cases. Yup, there are still a lot of hurdles to go, and a lot of situations that cars still can't handle, but if we look at the experiments that Google and Uber have been doing, as well as what Tesla has shown, I am pretty convinced that you could release a self-driving car today, without a controlled environment (though probably in a restricted environment, i.e. not during snowstorms and only on decent roads) and it would work. (note, this is also talking about autonomous cars in general, not specifically the Tesla).
Many people seem to be against that idea, which is reasonable in cities where there are significant capacity constraints already - but for highways, for reducing the dangers in torturous cross-country drives - it would be easy to provide this
I don't think it would be easy to do given the numbers involved, but its certainly doable, and its a MUCH cheaper way of increasing the capacity of a road than anything being discussed to date. IF anything, I suspect that cities would be easier to do something in. The harder issue would be... what to do? there is no current "Standard" (that I know of) for "marking" lanes, stop signs, exits, etc... for autonomous vehicles, or for broadcasting traffic light status, road work, etc... I can think of lots of different ways of doing these, but whatever way was picked would HAVE to be something that would be done in cooperation with government and industry so that all self-driving vehicles would be able to understand them, and (again, to my admittedly limited knowledge) there is no current project in place to even look at this as an option outside of some limited university studies
Having said that, I do think that we may (probably will) see dedicated highways or lanes for autonomous cars in 5-10 years in certain areas, once there are enough cars out there to justify them. I just don't think they will be ones with special hardware right away.
And I do think we will see some active roadways that have special "things" to help autonomous cars navigate faster and safer... but I very much doubt we will see these until at least 5-10 years after we see autonomous cars working and in common use, so, I would say that is 10-15 years out at least.
We'll see. My first 2 weeks with the AP1 system I paid about a 10K premium for on my CPO have been a real eye-opener to a field I totally believed was on a five-year path to realization. Those who are speculating without having tried current systems (Tesla or otherwise for TACC) should really give it a go and see if you are as bullish on realization dates.
Like I said, I'd rather be wrong - and have a 100% safe, automated driving experience sooner - just don't think it's happening.
I've driven the AP1 systems a little, and I will admit they are not ready for prime time... all the time, everywhere... but, they are version 1, we are now (almost) at version 2, let's see what that does, and at this rate, I expect we will be at version 4 of the hardware in 5 years... (and I can't even guess at what that will look like) and remember, it is not an all or nothing proposition... and we don't have to be at level 5 to see some major benefits and common use, and I don't expect (or need) 100% accuracy in 100% of situations for it to be used. I figure that Humans are probably at something like 95% accuracy in 95% of situations (and 5% in the remaining 5% of situations), if it can match that, I will be comfortable with it (as long as I have a decent understanding of where those limitations actually are!).
I did buy the full self-driving option for my Tesla, but honestly, I didn't believe Elon when he said we will have full self-driving by the end of 2017, I do think (or at least hope) that it will be at a point where it can significantly reduce my driving load in the next few years, even if I can't just get in, tell it where to go, and take a nap as I could with a full level 5 vehicle.
But, like you said, we will see! I DO THINK it's gonna happen
but to be fair, I don't KNOW that it will.