Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×

Two Legged Robot Sets Speed Record 149

An anonymous reader writes "Researchers in Germany and Scotland have made the fastest two-legged robot yet (for its size) called RunBot. It is controlled by a simple program that mimics the way neurons control reflexes in humans and other animals. From the article: 'We wanted to show that a very simple system with a simple neuronal controller could walk in a natural manner - and fast,' says Florentin Wörgötter, from the University of Göttingen in Germany. The story also features couple of videos of RunBot in action."
This discussion has been archived. No new comments can be posted.

Two Legged Robot Sets Speed Record

Comments Filter:
  • by klack ( 823307 ) on Wednesday April 05, 2006 @07:11PM (#15071736)
    How many leg-length per second can the "average" human achieve?
  • I've heard of feet per second, but legs?
    • I've heard of feet per second, but legs?

      replace the word "legs" with "about 27cm" (from the looks of the thing).

      Also, from the video (which is kind of silly), it looks like it's attached to a pole, which may mean it can't stand or walk on its own (certainly not without power or instruction, but maybe also not without physical guidance). I donno, neat toy at any rate.

      PS, I mirrored the video here [tenthousandpercent.com]. Yeah I'm taking it down in 24 hours.
    • > I've heard of feet per second, but legs?

      Imagine a beowulf cluster of [uky.edu]...

    • I'm glad I'm not the only one who thinks leg-lengths per second is a misleading measure. Sure, the shorter robot has a better leg-lengths per second, but it's legs are shorter, so it's actually going slower in regular units/second.

      And before someone says "but it has shorter legs, so you have to compensate", I will say this: Why not compensate for the fact that a longer leg means more torque is necessary to move it? All it takes is looking at humans: people with shorter legs tend to move their legs much f
  • Runnin' (Score:5, Funny)

    by filtur ( 724994 ) on Wednesday April 05, 2006 @07:14PM (#15071758) Homepage
    I'll feel safe as long as I can still outrun our robot overlords.
  • videos (Score:5, Informative)

    by mcguyver ( 589810 ) on Wednesday April 05, 2006 @07:15PM (#15071770) Homepage
    Two mpeg videos show RunBot (4.9M) walking at a steady speed [ijrr.org] and (15M) gradually learning to walk more rapidly [ijrr.org].

    Get'em while they're hot, ;)
  • by Giant Ape Skeleton ( 638834 ) on Wednesday April 05, 2006 @07:15PM (#15071774) Homepage
    ...I won't be impressed until a two-robot team wins a three-legged race.
  • First video (Score:5, Funny)

    by cejones ( 574416 ) on Wednesday April 05, 2006 @07:16PM (#15071782)
    I love how it busts right at the end of video 1. They should have a sensor on its head to sensor when the head touches the ground.. . That should trigger a Homer-esque "D'oh!"
  • How fast (Score:2, Insightful)

    by Anonymous Coward
    Slashdot posts a story saying somebody made a robot that runs really fast, why the heck doesn't it say how fast the thing runs in the summary? Eh? EH???
    • Slashdot posts a story saying somebody made a robot that runs really fast, why the heck doesn't it say how fast the thing runs in the summary? Eh? EH???

      If it was a Roland Piquepaille link, I'd say Google ads. But, it was probably just lazyness.
  • anyone have a video of the little guy in action?
  • RunBot currently walks around the edge of a circular room and is connected the centre of the room by a boom.
    Yeah, you see the boom in the videos.

    It can walk but can't maintain verticality? Is it there to stabilize it? That's pretty lame if they don't even have to worry about keeping its center of balance ... that's the hardest thing to figure out about fluid bipedal motion!
    • by rjstanford ( 69735 ) on Wednesday April 05, 2006 @07:27PM (#15071862) Homepage Journal
      It can walk but can't maintain verticality? Is it there to stabilize it? That's pretty lame if they don't even have to worry about keeping its center of balance ... that's the hardest thing to figure out about fluid bipedal motion!

      Probably just to stop the experiment needing a shitload of space without worrying about explicit turning. Of course, you could have just quoted the entire paragraph (two sentences) - bold face added:

      RunBot currently walks around the edge of a circular room and is connected the centre of the room by a boom. But Wörgötter plans to develop a freestanding version next, and thinks it should be straightforward because the boom has only a small influence on its ability to walk.

      Doesn't seem quite so problematic now, does it?
      • by Illserve ( 56215 ) on Wednesday April 05, 2006 @09:07PM (#15072437)
        But Wörgötter plans to develop a freestanding version next, and thinks it should be straightforward because the boom has only a small influence on its ability to walk.

        Well what do you expect him to say, that this approach is hopelessly limited?

        This isn't the first time we've seen great mobility from tethered robots, but somehow these guys never manage to produce the untethered version. Getting power and proper balancing to an untethered robot seems to be the critical stumbling block and I would be shocked if this one doesn't hit the same issues.
        • Balance is a very complex issue. These guys are just focusing on the walking motion itself, and they've found a simpler and more effective way to do that.

          And that's good, because let's face it: that big, shuffling, passive balance robot from Sony looks like crap.
    • It can walk but can't maintain verticality?

      I was trying to think of some +5 Funny about this made up word .. but found out it was real.

    • IANA roboticist, but after I RTFA, it seemed like the point was to implement neural networks into more complex aspects of robotics. It seemed to be demonstrating that coding a (relatively) simple machine learning system allowed the robot to develop even an better walking algorithm than those written entirely by humans. As such, it was successful even with the rod.
    • Nah. They've had stand-alone walking robots [mainstreettoys.com] for years. It's not that hard.
  • This is a cool little robot - a nice example of how much you can do with a simple design. However, I think it is a bit premature to declare that it has set any biped records, seeing as how it can't balance on two legs yet.

  • by Z1NG ( 953122 ) on Wednesday April 05, 2006 @07:22PM (#15071828)
    The two women who actually read /. are hoping scientist don't put so much effort into their robots finishing quickly when they produce a three-legged bot.
  • "but it may not scale up - but I'll be interested to see if it does." -- Russ Tedrake

    The approach may not scale up?

    Has the guy heard of humans? Ostriches? Kangaroos? Bipedal locomotion doesn't scale indefinitely (square-cube problem unfortunately), but biological approaches to controlling two-legged walking (and running and bounding) will definitely scale up to lots of "useful" sizes.

    And if he's criticizing their current algorithms... of course they won't control a robot 10x larger, but that's just being
  • This thing walks in a circle and is connected to a boom - it can't walk freely. All the legs have to worry about is front/back balance, and not side to side. Of course, making that obvious in the headline or summary would make the article seem much less interesting, and we couldn't have that, now could we?
    • Could be worse. On Digg, the title would be "All Humans Will Walk With Bionic Legs By 2007!!!!"
    • This thing walks in a circle and is connected to a boom - it can't walk freely. All the legs have to worry about is front/back balance, and not side to side. Of course, making that obvious in the headline or summary would make the article seem much less interesting, and we couldn't have that, now could we?

      Accepted. But that wasn't really the point of the research. If you RTFA and RTFWP ( white paper [stir.ac.uk]), these guys are more interested in neuroscience. So what they did was design a simple mechanical system
  • Remember kids... (Score:5, Insightful)

    by vertinox ( 846076 ) on Wednesday April 05, 2006 @07:26PM (#15071850)
    You don't have to be faster than the killer robot.
    You just need to be faster than that any other humans you happen to be with.
  • OMG that's so totally cheating!!!!1!11one!1!!
  • Hah! In the first video it sounds exactly like the wrong trousers.
  • Runbot (Score:4, Insightful)

    by zippthorne ( 748122 ) on Wednesday April 05, 2006 @07:38PM (#15071939) Journal
    Althought called "runbot" it's actually a speedwalkerbot. at least that's all the videos show. At no point are fewer than one foot on the ground.
    • Re:Runbot (Score:5, Funny)

      by MickLinux ( 579158 ) on Wednesday April 05, 2006 @08:10PM (#15072138) Journal
      Yeah, I know that the dynamics of running -- one leap after another -- are possibly more complicated than, and definitely different than the dynamics for walking. Nonetheless, I always used to like our cross-country club races (when I did run CC, and later when I'd watch my brother run) in the Shenandoah Valley. There were a few "English walkers" who would outwalk quite a few runners. That includes me. It's just a little embarassing to be struggling along, still leaping from leg to leg in that slow hobble that we call "long distance running", and have somebody breezily walk past you (same direction) and offer a little how-do-you-do.
      • You were using too much of your energy to push up, rather than forward. Your feet may leave the ground, the trick is to not have to let most of your weight change height too much in any given stride (up too high, you have to catch it on the way down. Down too low, gotta push it back up). Saves lots of energy.
        • Re:Runbot (Score:3, Interesting)

          by AGMW ( 594303 )
          You were using too much of your energy to push up, rather than forward. Your feet may leave the ground, the trick is to not have to let most of your weight change height too much in any given stride (up too high, you have to catch it on the way down. Down too low, gotta push it back up). Saves lots of energy.

          Tell that to the Kangaroos [trnmag.com]! They have one of the most efficient bidepal locomotion stratagies because as they land they stretch two massive tendons and store all the kinetic energy, which they then u

          • Re:Runbot (Score:3, Interesting)

            by PitaBred ( 632671 )
            The kangaroos are storing the energy on coming down, and releasing most of it coming back up. A human doesn't have the same kind of mechanical advantage, and relies on muscle power to do the up and down motion. Therefore, it's more efficient for a human to keep vertical motion to a minimum so that the majority of the energy can be devoted to forward locomotion. The achilles tendon does help with humans, but it's not to the same extent as in kangaroos. Besides, if you watch a kangaroo when it's just movi
  • by M0b1u5 ( 569472 ) on Wednesday April 05, 2006 @07:43PM (#15071976) Homepage
    Didn't Qurio claim to be the world's first "running" robot?

    I always thought Qurio was great: push him over and no matter what orientation he ends up in, he can get back to his feet again. I think it's not too far away that after they get up off the ground, they come over and slap you for pushing them over...

    As to the whole walking thing - it's a fascinating topic I think:

    1) walking is a controlled fall, the only thing preventing you from going face-first into the pavement is that next foot fall.

    2) Maximum cruising speed is attained in a single revolution. No other animal or engine can claim the same (AFAIK).
    • 2) Maximum cruising speed is attained in a single revolution. No other animal or engine can claim the same
      I'm really unclear how you can claim that human walking/running is achieved in a single revolution. Certainly, there is a huge curve of surge in the first revolution, but nothing as grandiose as you claim.
      • Not running: walking.

        1) stand still.
        2) begin walking.
        3) measure speed after single step.

        speed = cruising speed.

        ipso facto: 1 revolution = cruising speed.

        And in fact, it's more like HALF a revolution, because two steps is a full revolution.

        Other animals can claim to achieve the same effect for their walking gait, but none can claim the efficiencies of the human gait, in terms of calories expended per kilogram carried per kilometres traveled.

        (Only a fully laden 747 is more efficient than a man on a racing bic
  • 1 question (Score:3, Funny)

    by hurfy ( 735314 ) on Wednesday April 05, 2006 @07:46PM (#15071997)
    How is it a 2-legged robot?

    Without the boom it will fall over, nor could it walk.

    Of course it can't even STOP , that might be a minor issue for usefulness ;p

    Hurfy
    Fastest man on 2 legs...while running down the aisle of the airliner anyways...
  • Tethered? (Score:3, Interesting)

    by Pedrito ( 94783 ) on Wednesday April 05, 2006 @07:47PM (#15072003)
    Call be back when it's untethered, has arms, and can stand itself back up after it trips. Then I'll be impressed. Until then, it's no more impressive than a bot on wheels.
  • by MaXiMiUS ( 923393 ) <{maximius} {at} {gmail.com}> on Wednesday April 05, 2006 @07:48PM (#15072007)
    C:\DOS
    C:\DOS\RUN
    RUN DOS RUN
  • Amazing..! (Score:2, Interesting)

    And this is what Kevin Kelly says [kk.org] in his excellent book Out of Control (The New Biology of Machines, Social Systems, and the Economic World). Start with small and dumb machines, follow nature and gradually build up the complexity. The efforts of creating one machine which does it all is going to fail.
  • YES BUT (Score:2, Funny)

    by popetty ( 950274 )
    They need to change run, to dance. Also, they need to include that it was rocking out to Mr. Roboto. DOMO ARIGATO MR ROBOT *FUUEEEEEWWWW**FUUEEEEEWWWW*
  • by metroplex ( 883298 ) on Wednesday April 05, 2006 @08:18PM (#15072179) Homepage
    welcome the lack of any obvious "I for one welcome our new two-legged running robotic speed record breaker overlords" jokes so far. :P
  • Scary (Score:2, Insightful)

    by kraksmokr ( 216277 )
    That is teh creepiest freaking thing I have ever seen.
    • Obviously you didn't watch the video of the four-legged mule robot developed by DARPA. It looks like some took two emaciated humans and encased them head to head, bent over, inside a metal framework. Combine that with the noise it makes and I spent the entire time watching the video feeling pity for the new creation and wishing someone would allow me to end its torment.

      This just looks like a gumby stop-action reject.
  • ...will it run on Linux?

    In all seriousness though, in the video it looked like the robot wasn't balancing itself. Instead it was attached to an arm coming from the center of the circle the robot was walking in to keep it upright. I have to say that this will be a lot more impressive when they can pull this off without having to hold the robot's hand.
    • and how long will it be before somebody sues them for infringing their "method of forward locomotion" patent.
      • ...And in the news of the bizarre, a robot is suing the entire human race for infringing upon its patented method of forward locomotion. While the ruling on the case is still up in the air, officials are suggesting people purchase Segways or wheelchairs, just in case.
  • I am already getting bored with all these articles about robots, isn't there anything worthy going on in the IT world besides robots that can do human-like things?

    I mean, my problem is, slashdot is getting boring! I need something to read about while I'm at work! How am I gonna be productive like that? Oh, wait...
  • RunBot? Please, who came up with that? Shouldn't it be called GoBot?
  • by Kingrames ( 858416 ) on Wednesday April 05, 2006 @10:02PM (#15072720)
    But can they build him faster? stronger?
  • I'm curious why no one brought up other possible uses of this idea. While it is obviously far from perfected this might be a nice technology for the disabled or injured to have. Being able to appear to walk normally and move about would be really nice. I know that there have been a lot of injuries in Iraq involving legs and other movement impairing injuries.
    • While it is obviously far from perfected this might be a nice technology for the disabled or injured to have. Being able to appear to walk normally and move about would be really nice.

      And who is going to follow the disabled around proping them up with a stick? This thing can't walk. Wörgötter claims the boom maintaining vertical stability only has a small influence on the system. But I would wager that maintaining vertical stability is the most difficult thing to simulate in this sort of excersis

  • by wubboy ( 96276 ) on Wednesday April 05, 2006 @11:45PM (#15073180)
    Yes but can it moon walk? Though I am not sure I want to see moonwalking again anyway.
  • "RunBot currently walks around the edge of a circular room and is connected the centre of the room by a boom. But Wörgötter plans to develop a freestanding version next, and thinks it should be straightforward because the boom has only a small influence on its ability to walk."

    This whole thing is not impressive. That thing would never walk on it's own if it didn't have that arm holding it up. I think I've seen things like that made out of wood and sold at craft fairs...
    • My sentiments exactly. I look forward to a free-standing version that can use a simple pseudo-neural program to walk on two legs. I can see from the video that they have a long way to go before they can build a free-standing one. At the very least, it needs articulated ankles and hips.
  • Härter. Schneller. Geh ans Limit. Übertriff dich selbst. Lerne, deinen Arsch zu meistern.

"When the going gets tough, the tough get empirical." -- Jon Carroll

Working...