Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Transportation Networking Wireless Networking

Researchers Are Developing Ad Hoc Networks For Car-To-Car Data Exchange 126

Posted by Soulskill from the hey-buddy-your-blinker-is-on dept.
Lucas123 writes "Researchers are developing machine-to-machine (M2M) communication technology that allows cars to exchange data with each other, enabling vehicles to know what the cars all around them are doing, and perhaps, where they're going. Intel is working with National Taiwan University on M2M connectivity, an idea came from caravanning — an available, but-not-yet-deployed technology that uses direct line of site infrared (IR) and a range finder in order to automatically adjust the speed of cars so they can travel at a measured distance from each other. In other words, they're electronically tethered to one another. Now, imagine a group of cars traveling down the road together as an ad hoc network, each one aware of the location, any sudden actions or even the travel route of other vehicles as uploaded to the cloud from a GPS device. 'We're even imagining in the future cars would be able to ask other cars, "Hey, can I cut into your lane?" Then the other car would let you in,' said Jennifer Healey, a research scientist with Intel."
This discussion has been archived. No new comments can be posted.

Researchers Are Developing Ad Hoc Networks For Car-To-Car Data Exchange More

Researchers Are Developing Ad Hoc Networks For Car-To-Car Data Exchange

Comments Filter:

  • They have that already (Score:5, Funny)

    by Gothmolly ( 148874 ) on Friday May 10, 2013 @01:43PM (#43687025)

    It's called a "train".

    • Yeah, but those cars' interfaces are tightly coupled.

      Except for the the ones near my grandmother's first husband, who worked on trains. He was killed when two cars came together with him in between. Ever since, I've avoided bad interfaces.

    • Re:They have that already (Score:5, Funny)

      by dkleinsc ( 563838 ) on Friday May 10, 2013 @02:05PM (#43687267) Homepage

      I was thinking more of a different car-to-car data exchange mechanism, namely the extended middle finger.

    • It's called a "train".

      Nah, trains are old-school, hard-wired, and are limited to following a single track to a place miles from where you are ultimately going (unless you live/work next to a train stop).

      Personally, I would prefer an IFF system.... Can I get missiles with that?

      • Not too familiar with rail yards [wikipedia.org] or switching [wikipedia.org] in general, are you?

      • It's called a "train".

        Nah, trains are old-school, hard-wired, and are limited to following a single track to a place miles from where you are ultimately going (unless you live/work next to a train stop).

        Personally, I would prefer an IFF system.... Can I get missiles with that?

        Actually, this looks like a civilian version of a system that is already in use in warships.

        Missiles cost sigificantly more than cars.

    • Like a centipede

      • That would more literally translate into something like you having cars on your feet.

        This is a bit different.

    • Re: (Score:2)

      by JWW ( 79176 )

      Bullshit. One car can't leave a train and then proceed to a different location from the other cars, while the rest of the cars keep moving along.

      This is an enormous difference.

      Everyone who continually spouts on about mass transit always takes the first step of discounting how much flexibility and independence matter in transit.

      With this type of breakthrough we can get the benefits of the train (energy savings earned from the streamlining of the group of cars' movement), plus still retain flexibility (cars

  • Deck chairs arrangements (Score:1)

    by Anonymous Coward

    ... or we could just use mass transit.

    • ... or we could just use mass transit.

      It is more likely that self-driving cars will kill mass transit. People use mass transit so they don't have to own a car, and to avoid traffic congestion and parking fees. But self driving cars solve all those problems. Self driving taxis will be available on demand, they will eliminate congestion by using the road space far more efficiently, and you won't have to park them at your destination. They will offer conveniences that mass transit does not: they can meet you at your front door at at a time co

      • Re:Deck chairs arrangements (Score:4, Insightful)

        by lgw ( 121541 ) on Friday May 10, 2013 @03:30PM (#43688211) Journal

        Self driving taxis will be available on demand

        For any such system to be useful, you need a very strong system to prevent them being used as a place to sleep, convenient bathroom, and so on. Zipcar manages, mostly by not being public - you have a contract, a credit car on file, insurance, and so on. I can see Zipcar offering self-driving cars long before I can see a taxi service doing so.

        • being used as a place to sleep

          This can be prevented by debiting their credit card for the time that they are using the car.

          convenient bathroom

          This can be prevented by a camera monitoring the passenger compartment, and debiting their credit card for any cleanup fee. Many human operated taxis already have customer-cams. If you want privacy while you are peeing, then don't pee in the taxi.

          I can see Zipcar offering self-driving cars long before I can see a taxi service doing so.

          I can see the end result being the same either way.

      • I mostly agree with you, but I also think that self-driving taxis could also improve public transit usage because they solve the last mile problem -- getting to and from mass transit. For medium distance commutes, it might still be faster to take a self-driving taxi to the train/subway and then get another taxi after getting off the train/subway, for locations that aren't too far from a transit center.

        Also, it'd be easier to take the self-driving bus in the morning if I knew that I could call a self-driving

        • Re: (Score:2)

          by Reapy ( 688651 )

          Rip all our roads up and have a rail system, instead of cars we own our own rail car that sits in a garage that his hooked up to the main system. If I don't own a rail car or need a bigger one if I'm moving or have guests over or want a party bus style move, I can just rent one via the control panel and take all the risks of a dirty car, or I could rent one from a private company that charges a bit more.

          The car would be routed out to my house as soon as I place the order, or I might have scheduled it ahead

          • Re: (Score:2)

            by amorsen ( 7485 )

            What is the advantage of using steel-against-steel instead of rubber-against-concrete in this arrangement? Other than lower rolling resistance and the lousy brakes.

            • Re: (Score:2)

              by Reapy ( 688651 )

              Well no real thought on the most efficient method, but thinking that a rail or some other mechanism gives you more control over positioning and would seem less of a difficult problem to make sure everybody is in their lane or can transfer around and less opportunities for 'rogue' vehicles to get onto the system, in addition to just taking up less space on rural branches.

              Really just seems routing a bunch of cars on fixed rails is easier when you don't have to worry about steering and keeping them there with

              • Re: (Score:2)

                by amorsen ( 7485 )

                Making a car follow a line is fairly simple though. You can add sides to the lanes if you are worried that the cars will try to do their own thing. It has the major advantage of allowing autonomous turns. With steel-on-steel, you have to let the track do the switching, and that is a fairly slow process. It can work for street cars because they at least switch something with a capacity of 50+ passengers; waiting for the switch for almost every passenger sounds unworkable.

  • sharing travel location? (Score:4, Insightful)

    by Nadaka ( 224565 ) on Friday May 10, 2013 @01:47PM (#43687061)

    Hello! Privacy concern! I may not want everyone knowing i have the strip club programmed into my gps.

  • Hacking potential (Score:5, Interesting)

    by AmiMoJo ( 196126 ) * on Friday May 10, 2013 @01:49PM (#43687083) Homepage Journal

    I imagine mod chips that block other cars from your lane will appear quickly enough, but the potential for carnage if one were programmed to give other cars information designed to mislead them into danger can't be ignored. How would one car authenticate what another is saying?

    • And with widespread hacking comes additional regulation, the eventual future involves getting sidelined into the slow lane while the rich and connected glide by as if you are standing still.

      • Re: (Score:2)

        by lgw ( 121541 )

        the eventual future involves getting sidelined into the slow lane while the rich and connected glide by as if you are standing still.

        Now that's an idea I like! Sign me up.

        Already some congested cities are offering toll lanes, where you pay a few dollars to get less-slow traffic - if your time is worth it, pay up! I rather suspect these HOV-or-toll lanes will become HOV-or-toll-or-self-driving lanes, and you'll be the one gliding by.

    • I don't think you have to assume malicious behavior for this to be a real mess. The only way it would work properly is if every car on the road had one of these and they were all in working order. Shit, some people can't even be bothered to fix their headlights unless they get a ticket.
      • My thoughts exactly, this sounds great in theory until you realize that the driving populous is no where even remotely responsible or consistent enough for something like to work. Add deviants deliberately sabotaging the system in and this is just a pipe dream at best.

    • I imagine mod chips that block other cars from your lane will appear quickly enough, but the potential for carnage if one were programmed to give other cars information designed to mislead them into danger can't be ignored. How would one car authenticate what another is saying?

      That seems like the really intractable problem here...

      I don't mean to minimize the challenges of getting wireless mesh networks between a dynamic population of moving targets; but that's the sort of problem that falls into 'problem is hard, good thing our engineers are smart'. It'll get worked out.

      Coaxing optimal results out of swarms is similarly a problem that hasn't been fully explored(but we know that social insects are pretty good at it, and it's an active area of research).

      Add malicious actors to the

      • Re: (Score:2)

        by lgw ( 121541 )

        How is a malicious actor not going to get caught? I'd imagine any self-driving system will include cameras capable of recording the car/license plates of the cars it communicates with, as well as some digital identifier. Using some hack to mislead others leading to harm is going to get noticed very quickly, just like using a radar detector where they're illegal.

        That's a technical problem with a very well established social solution!

        • So you're assuming that communications will be limited to the eight cars nearest to you? Not the car four cars ahead of you or four cars behind you or the car on an overpass or the dude with the parabolic high gain antenna half a mile away in some high-rise? And no one will spoof a digital identifier?

          • Re: (Score:2)

            by lgw ( 121541 )

            The first two things a self-driving car needs to do, long before we get to any sort of networking, are:
            * Know where fixed obstacles are
            * Know where the other nearby cars are

            If it can't tell the difference between a car in front of it and a car on the overpass, you'll be dead long before you care about hackers. So, yes, I'm assuming that you'd use a communication technology that made antenna location very .. locatable, since that's the whole point of networking this stuff.

            And, yes, you could spoof a digital

          • Re: (Score:2)

            by amorsen ( 7485 )

            If you get a spoofed approval to change lane, you still have to wait for the car to actually move to make room. Since the approval was spoofed, that never happens and so the system just fails to change lane. It will likely try the indicator next, which is difficult to spoof.

            Obviously you could spoof something more annoying, like faking an obstacle. However, that will just lead to some overly enthusiastic braking and perhaps some fender benders (depending on how much space there is between cars). Since the s

      • Re: (Score:1)

        by MBasial ( 565610 )
        I am not particularly worried about the "malicious operator" scenario. Trick someone else's car into a fatal accident, and that's probably premeditated murder to a jury (or one of the heavier manslaughters, at least). Prosecutors will likely take it a bit more seriously than that time someone's computer got hacked to send out spam. Yes, it's possible to do, if you are also willing to do the prison time.

        • Sucks for the dead person though.

          FYI the vast majority of murders are not solved. They don't even have suspects.

    • I'd imagine that you would also have sensors on the car to confirm the information.

    • Re: (Score:2)

      by Idbar ( 1034346 )
      You're talking about a car-in-the-middle attack? Sounds like a possibility.

  • You'll have a mile long pack of cars, all accelerating as fast as the slowest vehicle.

    And I can only imagine the rolling roadblocks you'll get when a row of cars line up and synchronize their speeds.

    • Re: (Score:2)

      by Qzukk ( 229616 )

      And I can only imagine the rolling roadblocks you'll get when a row of cars line up and synchronize their speeds.

      Only a problem if you're a meatbag driver who thinks he's awesome enough to be allowed on the road with his sluggish 1 second response time.

      Or if everyone else hacks their car to answer "no" when your car asks to be let through.

    • NO, the road will sort the cars. All the inter-car comms will coordinate with the road signaling system to avoid this sort of thing.

    • But it won't matter, because the cars will also have algorithms to determine when to just safely move around a car that can't keep up, and that synchronized roadblock won't block your also-upgraded car from joining (or passing through) them.

      More irritating will be the Luddites who don't take advantage of the technology, so they create obstacles that the networks will have to route around. When that old geezer going 45mph is getting passed by that other old geezer going 46mph, the whole train of cars behind

      • Re: (Score:2)

        by amorsen ( 7485 )

        Luddites won't be a problem for long. Once accident rates start dropping, insurance premiums will go down as well. Almost everyone will switch to upgraded vehicles to get cheaper insurance. At the same time awards to victims of avoidable accidents due to insufficiently upgraded cars will likely go up, forcing premiums higher for the Luddites.

  • "Hey, can I cut into your lane?" Then the other car would let you in,' said Jennifer Healey, a research scientist with Intel." With CyanogenMod 20, your car will be able to flip a virtual finger and respond with a binary "FUCK OFF AND DIE!" message.

    • Re: (Score:1)

      by Anonymous Coward

      As long as we can tag those people, and get a warning when we're near asshats, that's not a problem.

  • Why Taiwan? (Score:3, Interesting)

    by jaymzter ( 452402 ) on Friday May 10, 2013 @01:56PM (#43687161) Homepage

    The California university system is larger than any in Taiwan and exists in Intel's home state. Anyone have any ideas why this research was offshored?

    After all (and on another note), this seems right up the governing regime of California's "alley". Imagine the state using this technology to mandate your speed, or taxing you for entering the city core during certain hours, not to mention the wonderful surveillance opportunities.

    • Intel's home state

      Multinationals have no "home". They spend money wherever it's best for the bottom line, full stop.

  • Want that nice car in the caravan? Hate somebody? See what you can do with some false signals injected into the M2M communications protocols.

    • Re: (Score:2)

      by lgw ( 121541 )

      Injected from where? From the illegally modded computer in your car? From an illegal radio emitter? Both of these are quite solvable problems.

  • There's more than a few edge conditions that I worry about - and that's without even thinking about malicious actors.

    Some edge conditions:

    • A big car transportation truck (double bottomed) with the car navigation systems left on.
    • Multiple, physically adjacent highways - with concrete barriers between them
    • Traffic stalls on multi-deck bridges
    • Bleed-over from service roads running parallel to highways.

    Of malicious actors, I can think of:

    • Black-hat/vandals leaving false transmitters on the side of the road
    • A big car transportation truck (double bottomed) with the car navigation systems left on. All would have speed readings of 0.
      Multiple, physically adjacent highways - with concrete barriers between them Line of sight
      Traffic stalls on multi-deck bridges Line of sight
      Bleed-over from service roads running parallel to highways. Line of sight

      I think these cars will end up using a combination of line-of-sight and bluetooth-ish comm, that establishes communication and location. The line of sight is handled

    • Re: (Score:2)

      by lgw ( 121541 )

      All of these seem like minor worries compared to similar existing issues.

      A big car transportation truck (double bottomed) with the car navigation systems left on.

      Those truck drivers are far more attentive and responsible on the whole than car drivers. But if you assume one isn't, why worry that he somehow left a car turned on when you could worry that he somehow left a car unsecured, and it will roll off the back of the truck at speed? That's a far more colorful scenario!

      Multiple, physically adjacent highways - with concrete barriers between them.

      If the self-driving car doesn't know quite clearly where nearby concrete barriers are, you have bigger issues!

      Traffic stalls on multi-deck bridges

      What would be

    • A big car transportation truck (double bottomed) with the car navigation systems left on.

      But the navigation systems won't have any destination programmed, nor will the engine be running. It's pretty easy to recognize "I'm moving but not trying to... something else is in charge"

      Multiple, physically adjacent highways - with concrete barriers between them

      If you're suggesting the cars would try to merge through the barrier, that's what lane tracking is for. If you're suggesting the cars would get confused as to where they are, it won't matter, because they're completely different roads to the routing system.

      Traffic stalls on multi-deck bridges

      ...won't matter, because they're different roads to the routing sy

  • ...whether or not they could, they didn't stop to think about if they should.

    Talk about a technology ripe for abuse/hacking/et cetera.

  • Hackable cars!! What could possibly go wrong? ;-)

  • It is called DSRC and it is nothing new (Score:5, Informative)

    by Bluefirebird ( 649667 ) on Friday May 10, 2013 @02:02PM (#43687245)
    This is old news!
    The Dedicated short-range communications (DSRC) is a set of protocols and standards for vehicle-to-vehicle and vehicle to infrastructure communications.
    The lower layers in the protocol stack are defined by the 802.11p standard, which is a modified version for the 802.11a for vehicular environments and it operates in the 5.9 GHz band.
    The higher layers are defined by the Wireless Access in Vehicular Environments (WAVE) stack, for messaging and control, and the IPv6 stack for applications and services.


    There are already commercial DSRC radios and lots applications have been developed in the ITS research community. For instance, the See-Through System: an overtaking assistance system http://www.youtube.com/watch?v=Esh1EjgBQaI [youtube.com]

    • Someone mod parent up please. That's the first thing I thought of when reading this. "Why are we re-inventing the wheel again, again?".

    • Why they're doing it is there's enough talk about intelligent cars now that there's room for "new technology" to get easy marketing. An existing standard doesn't matter when you can make your technology the de-facto standard by market manipulation.

    • Yup. I built and installed a working DSRC system in 2006. Supposedly, large-scale deployment was scheduled to start in 2008. Oops.
      The major problem was that the chief sparkplug for the whole idea got sick and died. Without him pushing the project, it stagnated.

  • One dirty sensor will muck that up fast. I'm horrible at washing my car. I wonder how forgiving the IR sensor's location will be to my bad habits.

  • to flip off the other cars? Important.

  • The protocol will have 300 different signal methods to tell the other car it's driver is an idiot. Then 99% of all use cases are covered.

    • The protocol will have 300 different signal methods to tell the other car it's driver is an idiot. Then 99% of all use cases are covered.

      Yeah, I was curious about this too...how will the car know if the driver is about to swerve wildly while texting or yelling at the kids in the back seat?

      I suppose they're just getting a mesh network protocol prepared for use with the long-anticipated auto-piloting vehicles...can't see this as very useful as long as a human is still controlling the gas, brakes and steering wheel. I suppose the other vehicles could theoretically cut in and assume crash avoidance maneuvers with faster reaction speeds than the

  • Why? (Score:5, Funny)

    by tech.kyle ( 2800087 ) on Friday May 10, 2013 @02:10PM (#43687321)
    Because cars should have twitter too.

    Car1: Sitting at a red light. So bored!
    Car2: @Car1 LOL floored it on the yellow and made it through.
    Car1: @Car2 At the next red light with you. Wanna race?
    Car3: @Car1 @Car2 Police ahead, don't do it.
    TotallyNotPolice: @Car1 @Car2 Ignore him. No police. You should race.

    • If we colonize Mars, it won't be the World Wide Web anymore. Ever think of that?

      Sure it will. World is self-defining. It will just be larger than it is now. With some hellacious latency in a few spots.

  • I had thought of this idea as a way to ticket those assholes on the road that drive like maniacs when there are no cops around. If 50 (PRESET_LIMIT) people report a car changing lanes without a signal in one day (PRESET_TIME) then a fine could be sent. Speeding, cutting people off, running red lights, etc could be crowd sourced to achieve better driving from the public. Of course I would not like the idea myself. It seems a little overboard or something. Plus, I'm sure the black-hat hackers would find ways

    • Re: (Score:1)

      by Anonymous Coward

      Gallagher invented this decades ago. Everyone gets a gun that shoots darts with suction cups with Idiot! flags. Any cop that sees a car with 3 flags can issue a ticket.

      You get 3 flags, need to go before a judge and explain why you used the first 3 and request more.

      Makes about as much sense...

  • Now to update the auto-targeting system for my turret gun...they can run, but they can't hide anymore!

  • This essentially sounds like ADS-B [wikipedia.org] for ground vehicles.

  • I've always wanted to be able to text the douchebag that just cut me off and tell him/her what a douchebag he/she is.

    Or be able to leave a nice little voicemail or text to tell the douchebag in the parking lot that he/she parked like a douchebag taking up two spots.

    Of course I'll get a few of those myself, but it might be worth it, just for the occasional satisfaction of calling out a douchebag.

  • 'We're even imagining in the future cars would be able to ask other cars, "Hey, can I cut into your lane?" Then the other car would let you in,'

    Ya, unless that other car is a jerk, then it will just speed up to close any gap.
    [ Will we be able to program a "personality" for our car? ]

  • Sample data exchange (Score:5, Funny)

    by RevWaldo ( 1186281 ) on Friday May 10, 2013 @03:37PM (#43688297)
    <?xml version="1.0" encoding="ISO-8859-1"?>
    <M2M>
    <vehicle>
    <ssid>rubber_duck</ssid>
    <haul>timber</price>
    <destination>Tulsa Town</destination>
    <kph>160</kph>
    <mode>convoy</mode>
    <memo>we got a great big convoy</memo>
    </vehicle>
    <vehicle>
    <ssid>big_ben</ssid>
    <haul>hogs</price>
    <destination>Tulsa Town</destination>
    <kph>160</kph>
    <mode>convoy</mode>
    <memo>ain\'t she a beautiful sight?</memo>
    </vehicle>
    </M2M>

  • Combining with Google driverless and Ad Hoc/Cloud computing, maybe this can improve city traffic.

  • Thank you Cory Doctorow [craphound.com].
  • I've long believed that ad hoc networks like this are mostly a fantasy because the potential for mischief as they become popular (ie, in the large) overrides their ability to do interesting things in the small. if you're not looking from the outset out how something can be gamed if it gets popular, you're part of the problem not part of the solution.

  • While this is an interesting variant, it faces the same problem that vehicle-2-vehicle communication based on the DSRC and 802.11p protocols does.

    Nobody has ever, as far as I know, built a network technology where you must network with random strangers you encounter out in the physical world. You can't build that because there is no value to the first people to install the tech, no value even to the first million in a country with 250 million cars like the USA. The odds of any 2 given cars being able to

  • I love to modify cars - the last thing I want is the police knowing I'm running a remapped ECU / upgraded turbo's because the cars capable of transmitting data about exhaust gasses and intake pressure. In Australia we get a $500 fine, EPA inspection and need to produce a valid Road Worthy Certificate if we modify our cars too much, even if your modifications make the car safer (bigger brakes, better suspension) or more efficient with reduced emissions.....

Slashdot Top Deals

It is easier to write an incorrect program than understand a correct one.

Close