Stories
Slash Boxes
Comments

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

Researchers Apply P2P Principles To Car Traffic

Posted by ScuttleMonkey on Fri Jan 09, 2009 06:19 PM
from the traffic-eagle-eye dept.
alphadogg writes to tell us that lessons learned from peer-to-peer networks are being applied to traffic systems in order to prevent jams. "Their Autonet plan would center around ad hoc networks of vehicles and roadside monitoring posts supported by 802.11 technology (the prototype uses 11b). The vehicles would essentially be the 'clients' in such a system and feature graphical user interfaces to pass along information to drivers. They're building the system to be able to handle data on thousands of traffic incidents and road conditions."
+ -
story

Related Stories

This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
 Full
 Abbreviated
 Hidden
More
Loading... please wait.
  • by EggyToast (858951) on Friday January 09 2009, @06:22PM (#26393511) Homepage
    Most of the trouble I've seen, and most of the frustration I encounter, is from badly-timed traffic lights. And many delays are the result of civil service rather than accidents. For example, intersections that have very long red-lights lead to more people trying to speed through the light, causing accidents in the first place.

    This technology may help people avoid problems once they occur, but it won't do squat to affect the root of many problems -- bad traffic planning. Without a good traffic plan, everything made to "fix" it is just a patch on top of a bad base.

    • And what happens when everybody is using the same algorithm and it tells them "go left to avoid jam" and everybody flocks in the same way? it seems a centralized monitoring would be more useful in this sense. But a lot of people don't like to be centrally monitored...
      • If everyone were using the same algorithm, it would likely compensate by sending you left with e.g. 70% chance and right with 30% chance, depending on the relative capacities of those routes. Of course there wouldn't be any guarantee that the drivers would listen to the recommendation but if too many people clogged up one route that information would soon filter back into the system.
      • If your in Los Angeles traffic and traffic signals are monitored by http://trafficinfo.lacity.org/html/atsac_1.html [lacity.org]
    • by stewbacca (1033764) on Friday January 09 2009, @06:38PM (#26393673)

      Most of the trouble I've seen, and most of the frustration I encounter, is from badly-timed traffic lights.

      The most I've seen is from the overwhelming number of dumbasses on the road. A traffic light engineer is totally limited by the absolute inability of the moron up front to step on the pedal on the right when the light turns green, then the guy after him, then the guy after him. Get off your damned phone and GO already.

      • by EggyToast (858951) on Friday January 09 2009, @06:52PM (#26393845) Homepage
        Morons are a problem, of course, but even they can be alleviated with better light planning. If the lights are short, people aren't going to figure "well, I've got 2 minutes to kill, might as well pull out the phone." They know the lights going to change and they won't be able to pull it out of their pocket in time.

        In my city, we've got a couple streets where you can hit all greens, saving yourself about 5 minutes for the entire stretch, if you speed about 7-9 mph. You get half yellows and the rest are green. So anyone who tries it thinks "shit, this really is the best way to drive down this stretch," which just leads to a different kind of moron. Yet, if the lights were set up the *other* direction, traffic could be regulated so that there was no advantage to going over the speed limit -- you'd simply be approaching a red light anyway, and someone going exactly 25 or 35 would hit the light right after it changes. The only people slowed would be speeders.

        There's a lot that cities can do to alleviate traffic problems, but it's not "popular" or particularly showy, so almost none of them do. Fiddling with traffic lights doesn't win elections.

        • That works for one-way traffic. But when you time the lights so that one side can't speed, what happens to the other side?

          In short, it's really a messy optimization problem. Good luck.

      • Re: (Score:3, Insightful)

        A traffic light engineer is totally limited by the absolute inability of the moron up front to step on the pedal on the right when the light turns green,

        And people have learned to not go immediately when the light turns green, because some asshat is still running the red light on the cross street.
        • In 99.9% of the cases people have learned to not go because they can't find the "S" key on their texting device. Besides, in Europe, their Red lights flash yellow before turning green to let you know you are about to go, and people still manage to go when it turns green without getting killed by people running red lights, so your argument doesn't make any sense.
          • Actually , i see just the opposite : people go when the light is still red , but is already red for the cars.

      • If I had mod points right now, I'd mod you up immediately for this.

        Nothing pisses me off more than sitting in a turn lane behind some jackass on their phone. The light goes green, they don't move, everyone starts honking, and by the time the aforementioned jackass realizes what's going on, half of the light time has expired and only 3 vehicles make it through the intersection before the light changes.

        There also seems to be a direct correlation between the shortness of the turn light/the length of tim
    • by mrchaotica (681592) * on Friday January 09 2009, @06:40PM (#26393689)

      As a civil engineering student, I took a course that (among other things) taught me how to design traffic signal timing. I learned two surprising things:

      1. how hard it is to time the lights to give all traffic movements an acceptable level of service (especially if you can't add new lanes), and
      2. how poorly designed some of the intersections around here are.

      I think the root problem is that good transportation engineers are few and far in between (probably because a lot of people who went into transportation did so because structural engineering was too hard).

      • Re: (Score:3, Interesting)

        Agreed. I was in San Francisco over the summer, and noticed that the majority of the lights were very short. I was there as a pedestrian, not a driver, but it seemed that all of the drivers were cool with the short lights. After being there a few days, it made sense -- if you miss a light, it's not a big deal because it'll be green again in about 15 seconds.

        As a side effect, all of the pedestrians went to the corners to cross, because it was easier to wait a short time to get a light compared to waitin

          • by snowraver1 (1052510) on Friday January 09 2009, @08:15PM (#26394607)
            Just floor it a lil bit

            Funny you say that. There was an earlier comment earlier that also said a similar thing. I live in Calgary, AB, and downtown along 4th ave the lights are all synched for about 10 blocks or so, but only if you speed. The limit is 50, but you have to go about 60 or you start hitting yellows.

            Why is this? Can there really be that many Traffic Controllers that screwed up on the calculation? Is it so that police can sit on a corner at night and catch speeders? I don't get it...
            • Toronto: if you start at the bottom of Bayview Ave and maintain exactly 79kph, you can make it to the 401 without stopping (unless you get caught in the speed trap at Lawrence)
      • If you want to see a horrible intersection/interchange, and you're ever in L.A., check out the Mulholland/Valley Circle on/off ramps to the 101 in Woodland Hills.

        Granted, they had almost no real estate to work with, since of the four "corners", one was the Motion Picture Hospital (in LA, you don't mess with Spielberg), one was Hidden Hills (*cough* rich people *cough*), and one was a shopping center. So they had to improvise.... badly

      • Timing traffic lights is indeed difficult. Often attempts to fix problems just end up causing an even worse problem elsewhere. When the Chiswick roundabout in West London had its lights changed from carefully designed fixed cycle lights to "intelligent" lights that adapted to traffic flows, they ended up causing tailbacks on both the A4 Eastbound and North Circular Southbound, because when traffic got heavy, the sensors detected no flow and gave both major roads short green cycles. There were also situation
  • by Cyberax (705495) on Friday January 09 2009, @06:25PM (#26393539)

    Words "network collision" are going to take a whole new meaning :)

  • by djupedal (584558) on Friday January 09 2009, @06:27PM (#26393557)
    Put the damn money into comprehensive public transportation!!
    • Re: (Score:3, Insightful)

      by Anonymous Coward

      As many times as it takes to figure out slashdot is the wrong place to say it. You need to bug the people in charge of the money.

      • Re: (Score:3, Insightful)

        As many times as it takes to figure out slashdot is the wrong place to say it. You need to bug the people in charge of the money.

        The people in charge of the money are elected officials, who are elected by (among others) the people on Slashdot. One person bugging the people in charge of the money won't do much good, you need lots of people to do the bugging for anything to happen. The way you get lots of people is by raising awareness of the issue with the public by doing things like posting on Slashdot and talking to your friends (who talk to their friends...). Discussing issues is public is how you get public support, which is how y

    • Put the damn money into comprehensive public transportation!!

      But the auto companies -- you know, the ones who just asked for a big bailout and got some of it -- spent millions and millions of dollars convincing you, the unwitting public, that public transportation is a bad -- a waste of government resources!

      And now you know why there's no good public transportation in most big U.S. cities, save a few.

      Full disclosure: I have -- in the past -- worked for two of the Detroit Three automakers.

      • spent millions and millions of dollars convincing you, the unwitting public, that public transportation is a bad -- a waste of government resources!

        I spent years riding public transportation. And you know what?

        GM didn't need to spend a dime to convince me that they suck.

        • > GM didn't need to spend a dime to convince me that they suck.

          GM spents tons of money a century ago, buying politicians and getting rid of public transportation so that they could sell buses...think about it.
            • I like to think that I am pretty objective, and you have presented one side of the argument. I agree with you and I'd like less noise, danger (when walking just around the block) and pollution, but cars are *extremely convenient.*

              Anywhere outside of the dense city, it is the quickest, cheapest (on a marginal cost in most cases, anyway) way to get anyplace. All major cities that I've lived in have smelly, gross transportation systems (Boston, NYC) as opposed to a nice, comfortable, climate-controlled vehic

    • Uh, this will likely lead to as comprehensive public transportation as you can get. There have been a number of writeups about this, not the least of which was mentioned on /. in this article [slashdot.org].

      This routing information being handed to drivers who manually take action is the first step into complete and total automation. With automation, transportation costs will plummet and it's not entirely impossible to envision adequate transportation being listed as a "human right" along side adequate shelter and fo
    • +1. Public transport is orders of magnitude cleaner and less wasteful than cars.

      But as for P2P applications, what we could use is a cool carpooling network. Switch on your nav system, tell it where you want to go and it will tell you to where and when someone else is going there - or vice versa.

    • Well, it's more an urban planning issue really. Where I live we have a "suburban sprawl" problem. I live in an area that has lots of light industrial (read: farms), residential and then the commercial that supports it. For all intents and purposes our entire county is an extended suburban area of a large(ish) city (Portland, OR). We have some self-supporting industry but we still don't produce enough to cover ourselves.

      Currently the Clark County (WA) public transit bus system loses a metric shit ton of mon

      • I don't understand why that has to result in a loss of money. Is it the cost of the bus driver or the cost of the gas and maintenance?

        If it's the latter then buses should be replaced by electric vans during periods of time during which usage is low.

        Actually, you could get rid of the driver too if you could automate the buses/vans. Surely the task of automating a vehicle is much easier when the route is fixed like it is with buses.

        If someone does all this, please make sure to spend a little of the resulting
      • First off.. its my money, I should be able to do with it what I want.

        Not entirely. It's a product of an economy that you've had a marginal affect on.

        Secondly.. why does government even take care of roads at all?

        Not profitable.

        Thirdly.. I don't even take the bus, why should I pay for it?

        Because your vehicle out on the roads is part of the reason people choose to take the bus.

      • [quote]Put the damn money into comprehensive public transportation!![/quote]

        First off.. its my money, I should be able to do with it what I want. Secondly.. why does government even take care of roads at all? Thirdly.. I don't even take the bus, why should I pay for it?

        Thats like saying, I don't go to school any more, so why should I pay for it?, just because you don't, doesn't mean that nobdy else does

  • ... that senior citizens cause bottlenecks on the internet just like in cars! :-)

    Lord knows my grandma on the internet is a disaster.
    • ... that senior citizens cause bottlenecks on the internet just like in cars! :-)

      Lord knows my grandma on the internet is a disaster.

      We know they do. Wheelchairs don't do well in tubes. There are always rescue crews trying to winch them out.

  • by j1mmy (43634) on Friday January 09 2009, @06:33PM (#26393629) Journal

    research on this sort of thing has been going on for almost two decades now. the increasing ubiquity of in-car nav systems, cellphones with gps, and other positioning and communications technologies helps to overcome the biggest hurdle: critical mass. this sort of system isn't useful if only a handful of cars have it.

    the other, and more difficult, part of this work is using this data in a way that can provide predictive travel information to drivers before that data becomes outdated. it's one thing to know about congestion on a road 10 minutes from your current location. it's better to know whether it's still going to be congested when you get there. models to do this sort of thing exist, but aren't (yet) fast or reliable enough to be used in real time.

    in urban areas, there's been an increasing push for taxis to be outfitted with gps transponders both as a political move, but also as a research tool and eventual mechanism for supporting real-time traffic data collection. taxis in major cities cover all the big and little streets, all over the place, all the time. they're perfect for fitting into a regional live traffic data collection system.

  • by sabt-pestnu (967671) on Friday January 09 2009, @06:48PM (#26393781)

    Getting sued by the Motor Vehicle Association of America for using P2P traffic control software and downloading copyrighted road blocks!

  • A lot of the unnecessary traffic delays at poorly regulated traffic lights could be completely circumvented by getting rid of lights and settings up roundabouts. Even through traffic slows down, it does not stop, and it automatically regulates itself.

    Roundabout takes very little time to get used to, and it presents a consistent interface to drivers. First time I saw them in Italy many years ago during a business trip, I instantly fell in love with them. Since then I've seen them all over Europe. I think mos

  • For several years now I've wondered if traffic planners ever talk to physicists, specifically with regard to fluid dynamics. As a non-physicist, natch -- though I did date a nice lady once who was a traffic planner in a large city -- it seems to me that this could produce some good ideas, perhaps a traffic version of PARC. Internet researchers are equally, if not more, qualified to pipe up here and I'm interested to see what they come up with.

  • and then transfer the pieces to your destination.

  • Isn't that the communications network for the Autobots? This whole traffic management plan is a Decepticon plot I tell you! It must be destroyed!

    • <sincerequestion>What critical infrastructure could you fashion from a few dozen/hundred moving wifi nodes?</sincerequestion>

      It doesn't seem like this is something that would leave a lot of potential for abuse, save for tricking someone into being late for work by routing them into a heavily trafficked area.

    • Old people are usually only a small problem - they're fairly easy to route around because they are generally following the rules and doing so slowly, without tailgating giving room for others to maneuver - i.e. they remain predictable. Young people who are ignoring rules such as pass-in-the-proper-lane-and-then-get-out-of-the-way, don't signal, race up behind others to squeeze into a lane, or like I saw tonight, signal to pas on the incorrect-side (50% good) and then yank it the other direction nearly caus