Please create an account to participate in the Slashdot moderation system

 


Forgot your password?
Close
typodupeerror
×
Robotics Build

Robot With Broken Leg Learns To Walk Again In Under 2 Minutes 69

Posted by Unknown Lamer from the unstoppable-robot-overlords dept.
KentuckyFC (1144503) writes When animals lose a limb, they learn to hobble remarkably quickly. And yet when robots damage a leg, they become completely incapacitated. That now looks set to the change thanks to a group of robotics engineers who have worked out how to dramatically accelerate the process of learning to walk again when a limb has become damaged. They've tested it on a hexapod robot which finds an efficient new gait in under two minutes (with video), and often faster, when a leg becomes damaged. The problem for robots is that the parameter space of potential gaits is vast. For a robot with six legs and 18 motors, the task of finding an efficient new gait boils down to a search through 36-dimensional space. That's why it usually takes so long. The new approach gets around this by doing much of this calculation in advance, before the robot gets injured. The solutions are then ordered according to the amount of time each leg remains in contact with the ground. That reduces the dimension of the problem from 36 to 6 and so makes it much easier for the robot to search. When a leg becomes damaged, the robot selects new gaits from those that minimize contact with the ground for the damaged limb. It compares several and then chooses the fastest. Voila! The resulting gaits are often innovative, for example, with the robot moving by springing forward. The new approach even found a solution should all the legs become damaged. In that case, the robot flips onto its back and inches forward on its "shoulders."
This discussion has been archived. No new comments can be posted.

Robot With Broken Leg Learns To Walk Again In Under 2 Minutes More

Robot With Broken Leg Learns To Walk Again In Under 2 Minutes

Comments Filter:

  • Great (Score:5, Funny)

    by Sockatume ( 732728 ) on Wednesday July 23, 2014 @09:15AM (#47515213)

    Crowbars won't save you now.

  • I can't wait until Skynet becomes self-aware.

  • Just Testing Code (Score:5, Insightful)

    by Anonymous Coward on Wednesday July 23, 2014 @09:38AM (#47515399)

    The "new" walking patterns are all pre-programmed. It's not learning, it's just running a few presets and seeing which results in the greatest forward speed. This has use, but I wouldn't throw around "learning" for this experiment. If a novel break comes along that the programmers have not planned for, the machine won't have a working behavior is it's data banks.

    • Kind of like us animals. After an injury we have limited options to choose from. Just seeing which "preset" works the best.

    • Re:Just Testing Code (Score:4, Insightful)

      by fahrbot-bot ( 874524 ) on Wednesday July 23, 2014 @11:58AM (#47516471)

      Well... It's leaning in the same sense that I did when I accidentally hit my thumb with a hammer and my grandfather said that I should try and not do that again - and that he had learned that solution himself in his younger days. Grandfathers are often helpful like that.

    • Re:Just Testing Code (Score:4, Insightful)

      by Dutchmaan ( 442553 ) on Wednesday July 23, 2014 @02:02PM (#47517391) Homepage

      If a novel break comes along that the programmers have not planned for, the machine won't have a working behavior is it's data banks.

      Just presenting an oversimplified argument, but how would that differ from what our DNA has programmed for us. When I see robots using code for whatever specific reason, what's really going through my mind are that these are just micro components of what will eventually be incorporated into a much larger more complex "organism" Think of robots these days as simple organisms, where the primary concerns are mostly locomotion and simple functionality.

      • Re: (Score:1)

        by Anonymous Coward

        I learned some interesting new things in differential calculus this morning. That's an example of learning where the result is completely unanticipated. You won't find special cases for using calculus techniques in my DNA, that's all external.

        • How do you know you don't posses "coding" to process information in a particular way and apply it in a new circumstance, because that sounds exactly like what you "learned' I'm sure your train of thought was "what if i do THIS" i.e. going through your list of available problem solving abilities.
  • Having taken an AI class in College, I'm wondering if this might be a real-world scenario where a self-learning neural net would come in handy....not only for the broken leg scenario, but because you can tell it to learn, the "Gait" can be optimized at run-time under any scenario.

    • Re: (Score:1)

      by Anonymous Coward

      switch to run gait when needing to move fast over harsh terrain, switch to wheels when on smooth terrain (my bot designs are usually quadrupeds with wheels instead of feet, which is more stable and less damaging to the ground then hoove style feet) , switch to uphill gait, switch to downhill gait, switch to parallel to incline gait, switch to creep gait, switch to silent walking gait, switch to damaged gait, use legs as another manipulator, crawl, etc.

      making bots that can function in any environment a huma

      • yeah, and a neural net can be programmed to handle the different gaits and learn to optimize while in the appropriate gait mode. I understand this would come at a cost (more computation and memory) but in today's world memory is so cheap and small. In that case you wouldn't have a 'damaged' gait, if damaged, upon usage of a gait the learning algorithm would adapt to each environment if necessary...

        Think of the uphill gait while damaged...in your scenario you'd pick either 'uphill' or 'damaged'....but with

      • Yes; we have learned much from the Daleks.

        Eventually these hexapods will realize that they are sometimes more efficient when flying. Watch out.

  • It is even faster when injured... (Score:5, Insightful)

    by earthman ( 12244 ) on Wednesday July 23, 2014 @09:51AM (#47515495)

    So the forward speed with all legs functioning is 0.25m/s, and with one leg broken it is 0.27m/s.

    Therefore, if a robot chases you, do NOT break its leg, because that only makes it chase you even faster!

  • Comment removed (Score:5, Funny)

    by account_deleted ( 4530225 ) on Wednesday July 23, 2014 @09:55AM (#47515531)
    Comment removed based on user account deletion

  • Fast Forward (Score:5, Interesting)

    by retech ( 1228598 ) on Wednesday July 23, 2014 @09:56AM (#47515545)
    There was a show in the 1980's from the CBC in Canada called "Fast Forward". Every week it focused on different tech innovations and where they'd go. One week they had learning robots. For the most part these were all simplistic things. But they had an "ant" that was about 2 feet long that was autonomous. The MIT crew that had created it realized that a centralized brain was just too big and power draining to build into the robot. So they had a system of sensors with rudimentary data and needs (leg=up, down, forward, backward, touching, not, moving or not, etc). If they shut it down it lost the memory of how to do anything it learned that day. They turned it on for the camera and it was a flailing ball of legs. Within 5 minutes it not only learned how to walk but circumvent objects, falls, danger. It still sticks out as amazing. Watching this video, I wonder what ever happened to that bot from nearly 30 yrs ago and wonder why does this spider seem to have actually gone back in time?

    • I do recall that episode and was going to write about it. It was late 80's. I will say that I don't recall it being 2 feet long and I don't recall it being connected to anything. But you are exactly correct about the test. they disabled it somehow, and over the course of 5 minutes it was walking again and running some sort of search pattern.
      But now I don't recall if they disabled a leg or not.

      This brings on a side point:
      I do recall a study about repetitive science and lab work (also coding): their study sho

      • Re: (Score:2)

        by retech ( 1228598 )
        I'm guessing on the 2 feet. But based on the room, chairs, table, etc. I think between 12 - 24 inches is the range.

        It was one of the few untethered machines they had in the episode and I thought it amazing because it had no centralized intelligence. It had a core command set that told it to "walk". Or more according to them, gave it the desire to walk. How it did that was up to it.

        They disabled it by flipping a switch on the back that shut it down. And it was stunning to watch it learn to stumble then

        • a bit of research and I found
          http://www.cs.cmu.edu/~chuck/r... [cmu.edu]

          used this search on google.... MIT robot ants walking

          came up with this

          http://webcache.googleusercont... [googleusercontent.com]

          history repeating itself LOL

          • Re: (Score:2)

            by retech ( 1228598 )
            Both are cool. But the FF episode would have had to have been from 1980-81. The memory we're talking about for me, takes place in my bedroom which would put it in that time frame. I know FF ran previous and post this time period. But I'm relatively certain the ant robot episode ran from around that time. Also recall it having a posterior much like an ant that was solely for the batteries. While the head had a few limited sensors and the main controller was in the abdomen.

    • Re: (Score:2)

      by jandrese ( 485 )
      This reminds me of a story I read [washingtonpost.com] a few years ago about a landmine clearing robot that was designed to have its legs blown off and still travel through the minefield.

  • Under two minutes of bullshit (Score:5, Insightful)

    by Wrath0fb0b ( 302444 ) on Wednesday July 23, 2014 @10:26AM (#47515771)

    Hey, I can make all kinds of tasks faster by precomputing much of the work and then looking it up in a table. Congratulations, you've (re)discovered another instance of a Space/Time tradeoff [wikipedia.org].

    Now, in particular what they've done is still wicked cool -- it's a great idea to perform may millions of simulations ahead of time so that at runtime (heh) you can quickly draw on that data to adapt. It would be perfectly good research even without the over-the-top claim that they've somehow made the work faster as opposed to cleverly pre-computing much of it.

    But that's research -- you do something neat and then you make a ridiculous overstatement to generate buzz ...

    • What's this search nonsense ? It has pre-computed solutions for a number of scenarios, why not just look up the scenario based on the detected failure directly.

  • ... had better get her ass moving!

  • Controlling Damaged Aircraft (Score:5, Informative)

    by Anonymous Coward on Wednesday July 23, 2014 @10:28AM (#47515799)

    I recall about a decade ago the Department of Defense or NASA working on this sort of adaptability for controlling damaged aircraft.

    An aircraft that suddenly loses use of a part of a wing or a flight control surface may still have sufficient flight control capability to fly home. The problem generally is that the pilot's control inputs won't produce the same motion responses, and the pilot generally has only seconds to map the inputs to control outputs.

    The idea was for the computer to do this mapping for the pilot, so the pilot would continue to apply the appropriate inputs (to roll the plane for example). The computer would determine which of the remaining flight surfaces to employ in order to best achieve the desired motion.

    One example that I recall was when the aircraft rudder was lost, yaw motion was compensated by dropping the landing gear and speed brakes on only one side of the aircraft to cause more drag on only one side, yawing the aircraft.

  • Is this really a breakthrough? (Score:4, Insightful)

    by iONiUM ( 530420 ) on Wednesday July 23, 2014 @10:29AM (#47515805) Journal

    Against the Slashdot rules, I read TFA and watched the entire video.

    Unless I'm mistaken, all they did was create a giant array of possible motor combinations for movement, and then the robot just randomly tries them until it finds one which lets it more-or-less go in the same direction. It may not be the best one, but one that mostly works (it just stops at the first one that mostly works).

    Is that really a super big breakthrough? If the robot dynamically adapted to the broken leg, and figured out how to move using some semi-intelligent algorithm, I would say that is really awesome. But this is literally just trial and error through pre-created movement specs, randomly, then just selecting one that is mostly okay.

    Not trying to downplay other's achievements or research or anything, but it just doesn't seem like a big break through, unless "brute force" is something novel.

  • As far as I understood the article, everything is based on a behavioral repertoire... The only advancement of the study would be the confidence mapping of said repertoire? Wouldn't it be better to work toward the automatic creation of this repertoire by the robot itself?

  • As the Black Knight says... (Score:3)

    by alexhs ( 877055 ) on Wednesday July 23, 2014 @10:39AM (#47515881) Homepage Journal

    When animals lose a limb, they learn to hobble remarkably quickly.

    Right, I'll do you for that!
    It's just a flesh wound.

  • How long to prevent it? (Score:4, Funny)

    by T.E.D. ( 34228 ) on Wednesday July 23, 2014 @11:15AM (#47516167)
    How long will it take that robot to figure out how to stop the researchers from breaking its legs in the first place?
  • Stop all A.I. development ! All we need is to simulate the entire universe posibilities in a database. Then focus on designing a database bigger than the universe itself tu put it on every computer.

Slashdot Top Deals

No spitting on the Bus! Thank you, The Mgt.

Close