Bert, I hate to rehash, but what is one particular inability that will prevent autonomous cars from driving safely without changing the physical infrastructure? Not a long list. Your best example. And what particular infrastructure change would you recommend for that particular inability?
If any changes to the infrastructure are required, autonomous cars will take a lot longer to be realized and cost a lot more. So each claim that a particular infrastructure change is required should be extensively examined to see if there's a way around it.
I'm trying to avoid your argument that using street signs is v2i, therefore v2i is required.
"Bert, you may call it over-reaction on the part of consumers, but that anixiety issue is real, and it should not be underestimated."
I wouldn't say "underestimated," as much as I'd say "misplaced." The anxiety comes from unrealistic expectations of this driver-assisted operation.
As someone else commented, there is no way that a bunch of different driver assist algorithms are going to be capable of supporting truly autonomous driving. I ask again, why do street signs, traffic signals, pavement makings, and all the rest, exist today? Answer: because you can't have a bunch of independent drivers, each with his/her own styles, reflexes, and skill level, drive on the same infrastructures safely WITHOUT our existing "V2I communications."
Replace this diverse groups of drivers with an equally diverse, independently developed, driver-assist algorithms. The situation is no different. Hence, EITHER we develop these V2I comms, OR driver-assisted vehicles will not be fully autonomous, and the driver will have to be engaged in the process all of the time.
Since V2I requires all manner of politician input, I don't see it happening to the extent we need, anytime soon. I realize the automakers like to minimize that, because they need to hype up their products with big promises. What these driver assist systems will do is improve safety, just like ABS and yaw control can do. Put that Google car cold, unplanned, on a twisty mountain road, and see how well it works.
One of the most interesting things -- although it is obvious -- I heard during the session is when the Continental executive touched upon the issue of the approval of a vehicle, as the toughest challenge ahead for autonomous cars.
Indeed. Think about millions of different driving scenarios.
Bert, you may call it over-reaction on the part of consumers, but that anixiety issue is real, and it should not be underestimated. The automotive industry still has a long way to go; it needs to test, validate and prove technology for the very idea of self-driving to be accepted in a society -- in our generation.
Once we have a generation of young people who have been brought up in self-driving cars it seems only logical that they will be afraid to drive manually, especially once the cars evolve to truly being safer drivers than people
That's an interesting futuristic notion.
I've also heard people talking about whether driving a self-driving car requires a driver's license. Again, even if that won't become a reality any time soon, it's an interesting notion.
Blog Doing Math in FPGAs Tom Burke 24 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...