Slashdot Mirror


Cooperative Cars Battle It Out In Holland

An anonymous reader writes "The first cooperative platooning competition, where vehicles use radio communication in addition to sensors, was held in Helmond, Holland a week ago. By using wireless communication the awareness range of each vehicle is extended, enabling vehicles to travel closer together which increases road capacity while at the same time avoiding the shockwave effects responsible for traffic jams. The Grand Cooperative Driving Challenge distinguishes itself from earlier platooning demos (e.g. the PATH project) by having a completely heterogeneous mix of vehicles and systems built by multiple researcher and student teams. Using wireless communication to coordinate vehicles raises concerns about the safety of such systems, would you trust WiFi to drive your car?"

29 of 139 comments (clear)

  1. Just because the first car drives off a cliff, by fotoguzzi · · Score: 2

    does that mean all the others have to follow?

    --
    Their they're doing there hair.
  2. Of course yes! by Sique · · Score: 4, Informative

    I would trust WiFi more than the tired trucker or the drunk driver in the other lane.

    --
    .sig: Sique *sigh*
    1. Re:Of course yes! by somersault · · Score: 2

      You mean: things get safer, and leave you free to do something other than the job that a sedated monkey could be doing?

      --
      which is totally what she said
  3. 'bout time! by fezzzz · · Score: 2

    Everyday I travel by car, I feel this frustration that the car still needs me. Having to stop at traffic lights as the cars aren't synchronized and worrying that I might be distracted when the car in front of me brakes suddenly are only two of my gripes with driving the car myself.

    1. Re:'bout time! by TheRaven64 · · Score: 3, Informative

      It's still used, but it's difficult to get right. The problem is that traffic lights are not just random obstructions on a road, as they are in the picture in the Wikipedia article, they are used for junctions or pedestrian crossings. In both cases, you can often avoid the light turning red at all if there are no people waiting to cross the road, but that breaks the wave at the next set of traffic lights. If it's a junction, then you have a bigger synchronisation problem, because there are multiple independent paths between two sets of lights, and defining a wave in one segment may decrease the overall efficiency.

      When I was bored a few years ago, I wrote some code to try to define the optimal traffic light timings for a portion of Salt Lake City (where I was at the time - it has a very regular grid pattern, which makes it easy to model) to maximise total throughput. The results were quite counter intuitive (and very different to the traffic light timings that they were using).

      --
      I am TheRaven on Soylent News
    2. Re:'bout time! by bughunter · · Score: 2

      Whatever happened to the 'Green Wave'?

      Sensors. Roadway sensors happened.

      Unfortunately, though, the signal controller is only aware of the sensor states at its given intersection, so in practice you have two results worse than simple sync'ed lights: a) whole lines of dozens of cars stopping and accelerating at each intersection due to the influence of single cars tripping sensors with low latency, and b) single cars stopped at intersections with high latency for multiple minutes where no cross traffic occurs. (Latency being the time it takes for the light to change after you roll up on the sensor.)

      Why don't we have distributed sensor networks yet? It's not like they're frickin' flying cars. How much fuel would we save and pollution would we prevent if we eliminated the first result, a) above? Every day on my commute I see fifty cars forced to stop at every light, many times for just one car on a cross street that has just arrived at the intersection, and then accelerate when it turns green. And I also sit for 4 or 5 minutes at lights that are red where there is no cross traffic for intervals 30 seconds or more, every day; if it were a stop sign I could cross easily, but to cross against a red would be illegal.

      Both events are routine on every pass thru the system. Clearly it's the norm. And it's a huge waste of fuel, and clean air.

      CPU cycles are cheap. The kind of bandwidth necessary is cheap. Why can't we build a network of these sensors and have them manage groups of traffic thru the system? And let them allow single cars to cross the main roads in the gaps? As an engineer, it's very frustrating to drive in these systems, knowing that it could be made so much better by the attention of just one smart person in city government.

      --
      I can see the fnords!
  4. Integrating with reality by mwvdlee · · Score: 2

    How well do these cars cope with human-driven cars thrown in?

    --
    Slashdot social media options: AIM, ICQ, Yahoo, Jabber and Mobile Text. Why no MySpace?
    1. Re:Integrating with reality by GeniusDex · · Score: 2

      The whole idea of this system is that it can be slowly phased in. The system looks at other cars with the system nearby and looks at their behavior. If a car somewhat in front of you is breaking, a signal is sent to the driver that something is about to happen. It is not really autonomous right now, but supporting the driver.

  5. Safety is relative by Zouden · · Score: 2

    How about this:

    Using coloured lights and human eyesight to coordinate vehicles raises concerns about the safety of such systems, would you trust Joe Sixpack to drive your car?

    Humans are fallible, and hundreds die on the road every day. Would we accept a computer system that causes hundreds of people to die? Of course not. So any computer system that's considered capable of driving a car will almost certainly be safer than a human driver. Probably a thousand times safer.

    --
    "A week in the lab saves an hour in the library"
  6. Holland? by Zedrick · · Score: 2

    Helmond is in North Brabant, not Holland. Both are provinces in The Netherlands.

  7. Re:so who do you blame? by ledow · · Score: 4, Informative

    Answer: Lots and lots of money spent on legal cases with uncertain outcomes.

    This is part of the reason why people say we should have one road for human drivers and one for automated (which makes them so prohibitively expensive, it's not worth it). Basically if there's an accident, the human "driver" of the vehicle is responsible, whether he was on cruise control or his ABS failed or whatever. You can still have that but with automated cars, I foresee instant-law-suit as soon as something like that happens (in the style of the Toyota lawsuits) blaming the car.

    And on an all-automated road, if you have an accident then it's *GOT* to be the automation fault, right? So you think that the car companies and road companies are going to pick up the tab for the first 50-car multiple pile-up? What about the associated traffic delays for a thousand people driving their automated cars just behind? Again, it gets prohibitively expensive and risky for the car/road companies to operate.

    If you have an automated car on a "human" road, then the human has to be able to take over (seeing as he is the one responsible in case of a crash!), so it becomes a little bit like cruise control and also becomes 100% the driver's problem, even if the automation fails.

    More interesting - can you get arrested for something like "driving without due care and attention" if you're the driver of an automated vehicle and do something behind the wheel? If so (and current laws say "YES!"), you might as well just drive the damn thing yourself.

    It's pretty much why these things are university projects and not actually on the road except in "tests" (and also things like the demonstration of two "crash-proof automated Volvo's a couple of months ago that, when aimed at each other head on at 30mph were supposed to stop before any possible accident - in front of the press they crashed about a dozen times and stopped once).

    We've had the capability to remote-control and computer control a car for YEARS. Hell, we do it with aeroplanes and oil-tankers. But the fact of the matter is that we ALWAYS have a responsible human behind the wheel with the control to take over and, if they take their eyes off the controls, are deemed to be irresponsible (imagine if your airline pilot and his co-pilot both went to sleep and left it on auto?). The problem is that the law, economics and common-sense tell us it's a stupid thing to do.

    You want an automated vehicle? Get on the London Docklands Light Railway. Entirely driver-less. But they had to put conductors on the trains to reassure passengers because occasionally the things get stuck and go wrong even though they are on rails. "a Passenger Service Agent (PSA), originally referred to as a "Train Captain", on each train is responsible for patrolling the train, checking tickets, making announcements and controlling the doors. PSAs can also take control of the train in certain circumstances including equipment failure and emergencies." Been in operation since 1987, can only travel on the rails, can't go past their stated safe speed, and you can have actual physical objects on the rails that activate brakes to avoid collisions and STILL they have a "driver".

    Automated cars are like the "flying cars" of science fiction - yeah, it'd be cool, and we probably have the technology - but do you really want joy-riders flying over your house?

  8. Re:so who do you blame? by Eivind · · Score: 3, Interesting

    Sort of right. But when the benefits are large enough and obvious enough, a way is found (by changing law, if need be).

    Self-driving cars are significantly awesomer than normal cars, and I strongly suspect that their advantage is sufficient to force the necessary changes.

    Imagine what self-driving cars would do to DUI, to child-delivery, to parking-problems, to taxi-prices, to overnight long-distance driving, to commutes, to airport-parking-prices, to accidents-from-tiredness, to congestion.

    What will happen is -some- place will allow them, and shortly thereafter people elsewhere will demand that they be allowed, with sufficient force that they will be. (and in this case, industry is on the same side: the car-industry wants to sell these, at a significant premium initially offcourse)

  9. drivers by Tom · · Score: 3, Informative

    would you trust WiFi to drive your car?

    Do I trust the drivers of the other cars?

    Cars are these strange things that drive our minds crazy. I don't know how much is cultural (i.e. movies, etc.) and how much is psychological, but there are few areas in life where the disconnect between reality and subjective is so dramatic.

    Everyone thinks he's an above-average driver. Of course, that's statistically impossible.
    Almost everyone overestimates his (or her) ability to handle a car in unusual circumstances.
    Very few people can correctly judge road and weather conditions and their impacts on things like brake distance.
    Most people do not have a correct sense of speed anymore if they've driven at speed for a few hours.

    and so on and so forth. Car accidents are within the top reasons of unnatural death in most western countries, but most of us feel more uneasy going on a rollercoaster (which cause what, a dozen or so deaths a year, world-wide?) or on a plane (around 1000 deaths per year, world-wide) than taking the car to work (1,200,000 deaths per year, world-wide). Yes, that's the real numbers, here and here are some sources, or google your own. Plane crashes fall way below the rounding error margin of car crashes.

    Really, you would have to put really bad engineers with pre-historic computer equipment and unstable wiring into those cars to make them worse than human drivers.

    --
    Assorted stuff I do sometimes: Lemuria.org
    1. Re:drivers by arkenian · · Score: 2
      Quite. Several of them actually make it worse. Please note: In several occasions we've had troops in ACTIVE COMBAT OPERATIONS with lower fatality rates than the same number of troops, statistically, would have at home commuting. (The first Gulf War was estimated to have saved several hundred American lives, and there's been, I think, entire years of the current conflict where its been about break-even.) If you're not a soldier who participates in active combat operations, or have one of a very few other activities as a profession or hobby, getting in your car and commuting in a major metropolitan area is possibly the most dangerous thing you will ever do in your life by an order of magnitude. Maybe not, insofar as we've all probably done some very stupid things at some point in our lives, but certainly its the most dangerous thing you'll do on any sort of regular basis.

      I, for one, welcome our robot car-driving overlords. Can't come too soon.

  10. Re:so who do you blame? by ledow · · Score: 3, Informative

    Sell one car that will last until breakdown, require special roads, special taxation, special infrastructure, special laws, huge investment, extreme legal risk, having to ride around even-more-patents, having every politician in your pocket, etc.

    Or sell lots of cheaper cars that occasionally get dented/smashed up (but keep the driver intact of course), profit from the spare parts market (even if through patent licensing), require none of the above and where almost all the risk is on the driver.

    The car market is already over-priced and struggling (i.e. the ENTIRE UK car market had to be bailed out by the government just a few years ago, and it's not the first time). The governments already spend billions on road infrastructure (where a road is a bit of tarmac with some paint on it, not an isolated, obstruction-free, electronically-enabled, few-travellers, risky multi-billion-dollar venture) and, believe it or not, serious road accidents are actually rare given the number of cars in the road (multiply the number of air-accidents by the difference between the number of planes journeys and the number of cars journeys world-wide and see what happens!).

    Additionally, human drivers speeding and parking in the wrong places etc. is actually a HUGE source of income (not to mention drivers licenses, driving schools, insurance, etc.). Until the economics vastly change, it ain't gonna happen. If we see it in my lifetime, I will be hugely impressed at the amount of administrative and economic crap we've had to remove to get to that point. And to be honest, I don't particularly want it either.

  11. Re:so who do you blame? by CastrTroy · · Score: 2

    Three simple words. No fault insurance. In the event of an accident, your insurance covers your car and your injuries. There are no legal battles, because everyone is responsible for insuring themselves. They do this where I live. If you decide not to drive with insurance, then you aren't covered, regardless of whether or not the accident is your fault.

    --

    Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
  12. Re:so who do you blame? by ccguy · · Score: 3, Informative

    Well, for starters those sane laws you mention are usually just the traffic code, so those would apply equally. If a car doesn't yield when it has to then it doesn't make a difference if it was driven by a computer or a human driver.

    So probably no difference when it comes to other responsibility towards other drivers.

    Most interesting questions:

    - Would it be OK to be drunk in a fully computer driven car? (where the driver seat is just occupied by a passenger)
    - Would it be OK for someone without a driven license to use one of these cars?
    - In case of accident, assuming the computer was driving, do car owners take a hit in their driving license if they have one?
    - If the car is a rental or loan, how's the responsibility divided between car owner / insurance company / car driver / etc?
    - While we are at it, if cars are really able to drive themselves, do they actually need to have a human passenger at all? Can I send my car to my mom's to pick something up and come back?

    Anyway, obviously self driving cars would have a shitload of system getting data from external sensors, so it would actually be easier to find out exactly what happened in case of accident, particularly if more than one car is involved and you have two sets of data to compare.

    About mixing human and computer drivers, I'm not worried about it. I have no reason to believe that if the guy in the next lane is driving drunk and suddenly steers towards me I would have a better chance of solving it than a computer. I'd say the computer would actually react faster and with better control than I would. Sometimes accidents are inevitable by the way, and under some external circumstances there's no way at all to prevent them (even if you could replay the thing over and over). If I'm involved in one, I prefer to make sure its effects are minimized by a computer that knows what its doing.

  13. Re:Holland is... by sosume · · Score: 2

    Helmond is definitely NOT in Holland. Maybe in The Netherlands, but not in Holland.

  14. Re:The real question by mad_minstrel · · Score: 3, Insightful

    I don't know about the US but where I live there is such a thing as a periodic maintenance checkup. They could make sensor checks and software updates mandatory. And the software can always transmit its year/version along with the data so that our car can disregard any data from outdated systems. But it's all moot until there's law that says car makers are not responsible for any crashes the software causes - because they won't ever dare sell you an automated car otherwise.

    --
    May the source be with you.
  15. Re:Holland is... by sosume · · Score: 3, Insightful

    It's more like saying Belfast, England as Helmond does not belong to any of the Holland provinces.

  16. Re:Holland is... by JustOK · · Score: 2

    or at 42 26' 5" N / 83 59' 6" W, which is in Michigan.

    --
    rewriting history since 2109
  17. Re:I wouldn't trust humans to drive cars by instagib · · Score: 2

    http://en.wikipedia.org/wiki/List_of_countries_by_traffic-related_death_rate

    Interesting. Of the 5 countries with the lowest rates (ignoring islands and very small states), 2 drive on the left and 1 has highways without speed limit.

  18. Wikipedia article by macraig · · Score: 3, Insightful

    The Wikipedia article references a William J. Beatty:

    It has been said that by knowing how traffic waves are created, drivers can sometimes reduce their effects by increasing vehicle headways and reducing the use of brakes, ultimately alleviating traffic congestion for everyone in the area.

    I've been doing this as routine behavior for almost 30 years now, after observing these "waves" and theorizing the causes. I've been setting an example how to stop the waves (if not the jams altogether)... not that anyone recognizes the point of what I'm doing. Can't explain it to them! They just think I'm trying to piss them off, being lazy or not paying attention.

    That last is really why traffic jams occur, so taking the controls away from humans and giving it to machines that always pay attention, and thus know what to do and when to do it, is a good thing.

  19. Re:so who do you blame? by ledow · · Score: 2

    That was a (non-existent) "brake defect" where the drivers claimed they weren't able to brake. Not one case was proved in court where this was the case, to my knowledge. But, damn expensive for Toyota to prove otherwise, I should think.

    However, an ABS failure, for instance, wouldn't necessarily be the car's fault as the device only operates when the car is skidding anyway (read: driver error). And cruise-control is a human-activated switch that warns against it's use and that doesn't excuse you from controlling the car.

    So, yes, you may have an expensive proof. And the car company may have an even more expensive one. But it's hardly a get-out-of-jail-free card to the automated car manufacturer - they either assign the fault to the driver (and thus automated cars are worthless because the driver has to be 100% switched on, same as driving), assign it to themselves (and take on massive class-actions), or fight over who's to blame at great expense. All loss-situations for a company producing automated cars.

  20. I'd trust WiFi more than by RealGene · · Score: 2

    the woman in the Infiniti texting while driving in the left lane of the Mass Turnpike yesterday...

    --
    Mission: To provide products that consume time and energy as entertainingly as permitted by the laws of thermodynamics.
  21. Re:Talking of accidents by mehrotra.akash · · Score: 2

    So, I mention it, and an hour later there is an article on it??

    weird http://yro.slashdot.org/story/11/05/24/0159212/Mandatory-Automotive-Black-Boxes-May-Be-On-the-Way

  22. Re:so who do you blame? by Anonymous Coward · · Score: 2, Funny

    Well, why don't they simply NOT program the automated cars to crash?
    They can then avoid all these expensive liability lawsuits AND save developer time!

  23. Re:Wi-Fi? Seriously? by krilid · · Score: 2

    Looks like www.gcdc.net has been Slashdotted, here are a few links that might be interesting: (they actually use 5.9 GHz communications, 802.11p specifically for use in intelligent transportation systems) Movie from the preparations to the challenge, lots of nerds and tech: http://youtu.be/lmRifLzw8iA Winning team technical paper: http://www.mrt.kit.edu/annieway/downloads/gcdc11_team_description.pdf Runner up technical paper: http://www.hh.se/download/18.7a4c72f812fdb65798d80003243/IDE1120KLTechReportTeamPaper.pdf

  24. Re:so who do you blame? by Eivind · · Score: 2

    Taxis are neither cheap, nor -nearly- as awesome as you make them out to be, and for this reason a tiny fraction of human transport happens by taxi (especially outside city-centres).

    Taxis amount to using one human being and a car to transport another human being. (or more, but a large fraction of taxi-rides are single) If that human being is paid the same you are, then because of things like VAT and taxes, it ends up being significantly MORE expensive to let a taxi do it, than to do it yourself.

    You have to wait for taxis - if you want one at the same time many others do, you'll wait a lot. Not all taxi-drivers are dependable. Taxi-fares are expensive, especially in countries without a large poor-caste. Storing belongings, or items bought in one shop, while visiting another, is right out, and so on.