I think we are all getting ahead of ourselves by using the term "Automous Cars" casually. When reading some of the market research analysts' reports, I was confused; then, only by talking to them, i realized that there are so many shades of autonomy in autonomous cars!
Agree Junko, we need a definition of autonomous to start any discussion on this topic...I think the pragmatic, shorter term definition would be a car that can drive without driver involvolvement on highway (maybe autonomous highway needs to be defined too)...any other ideas? Kris
Hi, Kris. As you can see in the page 2 of this article, the US Department of Transportation's National Highway Traffic Safety Administration has laid out a pretty comprehensive set of five levels of autonomy for autonomous cars. And I think this is very useful.
What you described might belong to the Level 2, according to NHTSA's categorization -- or somewhere in between level 2 and level 3.
The issue is that where the sweet spot is going to be -- in terms of the market demand. Level 2 or level 3?
With the growing number of aging population, though, Level 3 (or Level 4) might be what might be really needed.
Let's walk before we run, getting something in place will be hard as these autonamous cars have to co-exist with "normal" ones for a while...I wonder at which point "normal" cars will be phased out, as analog TV today...nice article by the way Junko!
I wonder at which point "normal" cars will be phased out, as analog TV today.
Interesting you say that, Kris, because that's exactly what I was wondering when I was talking to one of the market analysts earlier today. When things like V2V (vehicle to vehicle) communication becomes a mandatory feature (more stories on this topic later) in new cars, obviously, the automotive industry will start looking for some sort of incentives for a small add-on device for "normal" cars, just like the TV industry was looking for the government's help for D-to-A down coverters.
How this will all play out is still too early to tell. But things are definitely getting interesting!
As a car guy, I admit that it is kind of depressing to think about fully autonomous cars. In spite of that, I will admit that as a Southern Californian, the thought of being able to sleep through my morning commute does have some appeal. As an engineer, I have to believe the data that shows the "Level 1" autonomous cars that we already have (IE, stability control) is demonstrably saving lives - so I guess automation may be a good thing, in that I'll be able to sleep safely to work.
@pinhead1, I think this is exactly one of the issues that carmakers are trying to wrestle with. On one hand, they want o add all the advanced driving assistance system to their cars (adding more value),helping to pave a way to "autonomous" cars. And yet, how BMW, the owner of "Ultimate driving machine" brand, would be able to differentiate their autonomous cars from others?
The debate on this -- how consumers will embrace it and how carmakers will pitch it -- has only begun.
I think that in fact, even Level 0 cars have a lot of autonomy built in. Most people today may not even be aware of it, unless they try to drive a Model T Ford perhaps.
Seems to me that fully autonomous cars are certainly possible, but first you have to create an infrastructure that is more predictable. But why is this hard? I've been on fully autonomous trains many times. No operator at all. The only reason they're feasible today is that the tracks make their path predictable. For that matter, same goes for airliners. They fly themselves most of the time, even for landing, and no "tracks" in sight. So why should we dismiss the possibility of making the car's relationship with roads more deterministic? You make it so cars run on predictable paths, when they operate in autonomous mode.
Just like with airliners, where navigation devices create the "tracks" they run on, you create these predicatble paths with vehicle to infrastructure comms. And you have emergency lanes where cars can stop in case of mechanical failures.
All I'm saying is, if we view this as an evolution of systems that already exist today (and they do exist today), then the rapid growth curve becomes believable.
As to what cars companies will boast about, it will be the creature comforts. Hey, things change. Engineers should know that better than anyone else. People used to boast about their horses, for pete's sake.
Blog Doing Math in FPGAs Tom Burke 15 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 ...