Author Topic: Self-Driving Cars  (Read 48013 times)

0 Members and 1 Guest are viewing this topic.

Offline PANTS!

  • One leg at a time.
  • Reef Tank Owner
  • *********
  • Posts: 9284
  • What seals? I auditioned for this job.
Re: Self-Driving Cars
« Reply #15 on: April 13, 2015, 01:39:59 PM »
I have another concern with self-driving cars: icy and snowy roads.  I can readily believe that current technology is capable of driving a car on a dry road, but driving on ice or in snow is a significantly different task, one which requires a fair amount of counter-intuitive action.  I have my doubts that the sensor technology used in such cars is capable of detecting and responding to winter road conditions.

Hmmm...  Interesting objection.  Wet roads too might be an issue.  However, there most certainly are traction control technologies available.  How / if they are integrated with the self-driving part is unknown to me.
Now where I come from
We don't let society tell us how it's supposed to be
-Uptown, Prince 👉

We cross our bridges when we come to them and burn them behind us, with nothing to show for our progress except a memory of the smell of smoke, and a presumption that once our eyes watered.

Offline amysrevenge

  • Baseball-Cap-Beard-Baby Guy
  • Stopped Going Outside
  • *******
  • Posts: 4978
  • The Warhammeriest
Re: Self-Driving Cars
« Reply #16 on: April 13, 2015, 01:45:40 PM »
I have another concern with self-driving cars: icy and snowy roads.  I can readily believe that current technology is capable of driving a car on a dry road, but driving on ice or in snow is a significantly different task, one which requires a fair amount of counter-intuitive action.  I have my doubts that the sensor technology used in such cars is capable of detecting and responding to winter road conditions.

Heh.  I've been assuming all along that when they say "self driving cars" they really mean "self driving cars in California".  With California as a stand-in for places where you can, for instance, count on having lines stay on the road without being obscured for months at a time by snow and ice, and scraped/corroded off by various snow removal procedures (I still remember marveling at the sticky-outie reflective thingies incorporated into the lane dividing lines when we moved to Seattle).  Places where you can count on 350+ days of driving impacted by nothing more onerous than a bit of rain in any given year.
Big Mike
Calgary AB Canada

Online The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Stopped Going Outside
  • *****
  • Posts: 4152
Re: Self-Driving Cars
« Reply #17 on: April 14, 2015, 11:43:02 AM »
Hmmm...  Interesting objection.  Wet roads too might be an issue.  However, there most certainly are traction control technologies available.  How / if they are integrated with the self-driving part is unknown to me.

In the winter months, I frequently drive on roads where there are no visible lines or pavement, and even no distinct edge to the road.  I also frequently drive when visibility is in the 20-30' range.  I have no idea how self-driving car would even keep its lane in such circumstances.  As for traction control, I have it; but it is not a panacea.  Indeed, there are times when one must disable it to drive effectively in snow or on ice.

How does a self-driving car in white-out conditions know what do do when its wheels start spinning?  Should it slow the engine, should it brake, or should it turn the wheels sharply into the spin and accelerate?  I don't think it does.
I would like to propose...that...it is undesirable to believe in a proposition when there is no ground whatever for supposing it true. — Bertrand Russell

Offline amysrevenge

  • Baseball-Cap-Beard-Baby Guy
  • Stopped Going Outside
  • *******
  • Posts: 4978
  • The Warhammeriest
Re: Self-Driving Cars
« Reply #18 on: April 14, 2015, 11:53:27 AM »
I don't think it's an insurmountable problem - I just think that self-driving in normal* conditions in not-California is at least an order of magnitude more difficult than self-driving in even abnormal conditions in California.

You just need the engineers to spend waaaay more winters in Canada or Wisconsin or what have you than they are likely to want to.  :)

*Where "normal" includes conditions such as The Latinist mentions above - these are not abnormal conditions, these are expected, regular, long-duration conditions.
Big Mike
Calgary AB Canada

Online superdave

  • 5000th post!
  • *******
  • Posts: 5000
  • My name is not dave.
Re: Self-Driving Cars
« Reply #19 on: April 14, 2015, 12:25:56 PM »
It's not like humans do so well in winter conditions either.  Why couldn't a computer do as good or better?

Offline amysrevenge

  • Baseball-Cap-Beard-Baby Guy
  • Stopped Going Outside
  • *******
  • Posts: 4978
  • The Warhammeriest
Re: Self-Driving Cars
« Reply #20 on: April 14, 2015, 12:34:28 PM »
The problem is one of detection and synthesis of external conditions.  The kind of things people are traditionally good at and computers haven't been as good at.

It's easy to, say, figure out where you are and how fast you're going on a proper road with visible lane demarcations and distinct edges and long visibility.  I got close to the math of it in an image processing graduate course a few years ago (it's a very similar problem to adding a virtual first down line to the TV display of a football game - we did work out that problem and it was pretty fun).

It is a lot harder when the visual cues are irregular and transient.  If you can add in deeper "visual" data with additional unhuman senses or whatever (radar, for instance), you can mitigate it. 

Order of magnitude stuff, not impossible stuff.  If you can solve it, then you're gonna do it better than humans (just like with California conditions).
Big Mike
Calgary AB Canada

Online The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Stopped Going Outside
  • *****
  • Posts: 4152
Re: Self-Driving Cars
« Reply #21 on: April 14, 2015, 01:30:10 PM »
It's not like humans do so well in winter conditions either.  Why couldn't a computer do as good or better?

Skilled winter drivers do quite well, thank you.  I'm not saying it's an impossible problem but, as amysrevenge said, it is orders of magnitude more difficult than driving a car on dry roads.  We're looking at cars being able to drive in ideal conditions quite soon, but I would not be at all surprised if the ability for a car to self-drive in winter conditions were decades away.
I would like to propose...that...it is undesirable to believe in a proposition when there is no ground whatever for supposing it true. — Bertrand Russell

Offline kvuo75

  • Well Established
  • *****
  • Posts: 1146
Re: Self-Driving Cars
« Reply #22 on: April 14, 2015, 07:16:25 PM »
Hmmm...  Interesting objection.  Wet roads too might be an issue.  However, there most certainly are traction control technologies available.  How / if they are integrated with the self-driving part is unknown to me.

In the winter months, I frequently drive on roads where there are no visible lines or pavement, and even no distinct edge to the road.  I also frequently drive when visibility is in the 20-30' range.  I have no idea how self-driving car would even keep its lane in such circumstances.  As for traction control, I have it; but it is not a panacea.  Indeed, there are times when one must disable it to drive effectively in snow or on ice.

How does a self-driving car in white-out conditions know what do do when its wheels start spinning?  Should it slow the engine, should it brake, or should it turn the wheels sharply into the spin and accelerate?  I don't think it does.

i'm pretty sure it can be figured out..

driving a car is not even rocket science.. and we can do rocket science.


Online The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Stopped Going Outside
  • *****
  • Posts: 4152
Re: Self-Driving Cars
« Reply #23 on: April 14, 2015, 09:05:28 PM »
Let me once again make very clear that I'm not saying cars that self-drive on winter roads can't be made.  I'm just not, and I wish people would stop responding as if I were.  What I'm saying is that I don't think it's anywhere near to being realized.  We already have cars driving independently on dry roads, and the tech's likely to come to consumer vehicles very soon; but I think we're at least 10 years and perhaps quite a bit longer away from a car that can drive in snow and ice.

Remember, in order for a car to be able to handle snow and ice effectively we're going to have to be able to create a computer abstraction of all the factors that go into the problem, something that to my knowledge nobody is even working on. Indeed, I don't think that we even have a basic theoretical understanding of the problems involved. I think everyone here except Ambious is significantly underestimating the difficulty of the problem.
I would like to propose...that...it is undesirable to believe in a proposition when there is no ground whatever for supposing it true. — Bertrand Russell

Offline kvuo75

  • Well Established
  • *****
  • Posts: 1146
Re: Self-Driving Cars
« Reply #24 on: April 14, 2015, 10:19:02 PM »
i think i understand it.. you're basically talking about different amounts of friction between tire and road surface.

even rudimentary ABS's already do that.. so we've pretty much got the stopping figured out.

the vehicle stability control system in my inexpensive car already uses vehicle speed, yaw, steering input, etc. to simply apply braking on one side or the other to correct for over/understeer (which is caused by human error to begin with).  i could only imagine if it took control of steering itself, how much more effective that would be.

as for being able to see a roadway -- i REALLY dont trust a person.. computers can use visibile light, ir, microwave to see roads AND traffic better than we can.. not to mention GPS and INS for positioning.

you're probably right that it's probably 10+ years away from reality, but nevertheless.  i think the technology already exists and i doubt ice and snow is much more a challenge than driving on wet roads...


edit:

we're also assuming the current self-driving cars can't do it already. why are we assuming that?
« Last Edit: April 14, 2015, 10:32:18 PM by kvuo75 »

Offline arthwollipot

  • Stopped Going Outside
  • *******
  • Posts: 5305
  • Observer of Phenomena
Re: Self-Driving Cars
« Reply #25 on: April 15, 2015, 02:20:28 AM »
I think an appropriately-programmed autonomous car will handle ice and snow better than most people.

Offline brilligtove

  • Stopped Going Outside
  • *******
  • Posts: 4613
  • Ignorance can be cured. Stupidity, you deal with.
    • Valuum
Re: Self-Driving Cars
« Reply #26 on: April 15, 2015, 08:54:33 AM »
This article mentions the radar and lasers used to see in snow and rain, but I could not find anything indicating how the most recent cross country test drive handled wet or snowy conditions.

If a self-driving car gets in an accident, who is to blame?  The person behind the wheel?  The software developer?  The vehicle manufacturer?  Until those problems are solved, almost definitely by legislative intervention, widescale adoption is unlikely.

Who get sued when a train derails?  When the solar powered monorail speeds out of control?  When an airplane crashes?

I mean I get that it hasn't been solved for cars, but I do not think it is too insurmountable.  There are transportation models we can draw upon now.  Seems to me like the adjusters and the police will need some training, and some new laws would need to be drafted.  But hell, if anything our current system of determining liability is far far more ambiguous when determining fault.  Not to mention there will be FAR fewer crashes, so in general each crash will be investigated more thoroughly.

I don't think these are equivalent to private drivers driving their privately owned vehicles. Or rather private drivers being driven in their privately owned vehicles. Can I be sued if my car kills your dog? Can the software developers be sued? Their employer?

I own the vehicle but I have no responsibility for how it behaves. How can I be sued when its behavior is murderous? If I owned a plane and hired a pilot to fly it and the pilot did something insane like crashed into a mountainside am I responsible for the pilots actions? In the case of an airline, the answer may be yes. Not because I have direct control of the pilot actions, but because am airline is regulated to take extraordinary measures to ensure that the pilots behaviour will remain in a safe zone. as the passenger and owner in an autonomous vehicle, I have no way of conforming to that kind of regulation so I cannot be held responsible. Except I chose to buy that car, just like I chose to get that dog, and when the dog bites a kid that's the end of that dog and its my fault. Of course in the case of a dog I have significantly more control over the "programming" of the animal.

I guess what I'm saying is I think we are entering new territory in terms of assigning blame and responsibility. It is going to be interesting.
evidence trumps experience | performance over perfection | responsibility – authority = scapegoat | emotions motivate; data doesn't

Online The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Stopped Going Outside
  • *****
  • Posts: 4152
Re: Self-Driving Cars
« Reply #27 on: April 15, 2015, 10:02:04 AM »
I think an appropriately-programmed autonomous car will handle ice and snow better than most people.

Sure.  And an appropriately-programmed computer will be able to write poetry better than most people.  The question is, how close are we to being able to create that programming?  Right now we have programming that can turn sports stats into news articles, but I don't think we're anywhere near a convincing poetry-writing computer.  I think the difference between driving on dry roads and driving on snowy and icy roads is on the same order of magnitude.
I would like to propose...that...it is undesirable to believe in a proposition when there is no ground whatever for supposing it true. — Bertrand Russell

Offline Johnny Slick

  • "Goddammit, Slick."
  • Poster of Extraordinary Magnitude
  • **********
  • Posts: 11990
  • Fake Ass Skeptic
Re: Self-Driving Cars
« Reply #28 on: April 15, 2015, 02:46:06 PM »
I sincerely doubt that an appropriately programmed computer will be able to write poetry well at all because that requires lateral thinking of the variety that computers are not good at.

With the self-driving cars issue, I can't help but think of the classic XKCD comic...



Driving strikes me as something that's stealthy-hard. You have to account for lots and lots of variables, and account for them so quickly that you don't consciously know that you accounted for them until you're after the fact. Figuring out what all of them are is going to be tough and I agree with TheLatinist that except in very specific instances (for instance, if you were to create a stretch of road that was fully automated for all vehicles involved), we're not particularly close to the point to where we've accounted for enough of said variables that it'll outdo humanity.
Speak what you think now in hard words, and to-morrow speak what to-morrow thinks in hard words again, though it contradict every thing you said to-day.

- Ralph Waldo Emerson

Online superdave

  • 5000th post!
  • *******
  • Posts: 5000
  • My name is not dave.
Re: Self-Driving Cars
« Reply #29 on: April 15, 2015, 05:06:05 PM »
I sincerely doubt that an appropriately programmed computer will be able to write poetry well at all because that requires lateral thinking of the variety that computers are not good at.

With the self-driving cars issue, I can't help but think of the classic XKCD comic...



Driving strikes me as something that's stealthy-hard. You have to account for lots and lots of variables, and account for them so quickly that you don't consciously know that you accounted for them until you're after the fact. Figuring out what all of them are is going to be tough and I agree with TheLatinist that except in very specific instances (for instance, if you were to create a stretch of road that was fully automated for all vehicles involved), we're not particularly close to the point to where we've accounted for enough of said variables that it'll outdo humanity.

The comic isn't wrong, but we really have come a long way in this field.  The IEEE has an entire weekly newsletter dedicated to this space.
http://spectrum.ieee.org/blog/cars-that-think