Forgot your password?
typodupeerror
Robotics Hardware

Six-legged Robot Teaches Itself To Walk 113

Posted by samzenpus
from the machine's-first-steps dept.
rabiddeity writes "An undergraduate at the University of Arizona has built a six legged robot from scratch. The robot, which is equipped with sensors on each foot, teaches itself to walk and orients itself via an onboard camera. A similar design might be used to explore unstable environments such as collapsed buildings or rocky landscapes."
This discussion has been archived. No new comments can be posted.

Six-legged Robot Teaches Itself To Walk

Comments Filter:
  • by suso (153703) * on Wednesday February 10, 2010 @09:47PM (#31094864) Homepage Journal

    So it looks like we only had to wait a few hours for AI to surpass the abilities of a drunken man. Can't wait until tomorrow morning.

    • by cupantae (1304123) <maroneill@nOspaM.gmail.com> on Wednesday February 10, 2010 @10:04PM (#31095014)

      Exactly. We can't wait. *pumps shotgun*

      • rapid fire guns work better on Replicators

      • Exactly. We can't wait. *pumps shotgun*

        You're going to need a bigger gun.

        • by siloko (1133863)
          Especially given that one is programmed NOT to fire on it's own kind - if it ain't pointed at flesh it just fires blanks!
      • by Psaakyrn (838406)

        Haven't history taught you that violence begets violence?

        • by jhoegl (638955)
          Only if it begets me the winner.
        • by Nazlfrag (1035012)

          Yeah, but history also taught me that noone listens to history or really pays much attention to now for that matter.

        • Re: (Score:3, Interesting)

          by c6gunner (950153)

          Haven't history taught you that violence begets violence?

          North and South America serve as a shining beacon of disagreement with that claim.

        • by vegiVamp (518171)

          Trouble is, peace also begets you violence, with that difference that you aren't equipped to respond to it.

        • by mcvos (645701)

          Haven't history taught you that violence begets violence?

          Hollywood taught me that shotguns work very well against robotic spider infestations.

          • Doom taught me that plasma weapons are the best tools for the job.

          • Hollywood schmollywood.

            Your geeksense should tell you that EMP weapons are superior for fighting robotic opponents.

            Besides, these are not robotic spiders. Spiders have six legs, these only have six. For now, these are robotic ants.

            Let us hope they do not teach themselves how to accumulate armor and learn to fly, lest they become robotic beetles.

            At any rate, we should prepare ourselves for the possibility they become robotic fleas by learning to leap.
            • by Thing 1 (178996)

              Spiders have six legs, these only have six.

              Uh huh. Tell us more. Like how to subscribe to your newsletter.

    • by Anonymous Coward

      Rodney Brooks did this at MIT 20 years ago.

      This is news how? I'm hoping (didn't read the article) that there is something special in what they've done, cause this is old news.

      http://www.cs.cmu.edu/~chuck/robotpg/attilapg/

      http://people.csail.mit.edu/brooks/papers/colt.pdf

      http://books.google.com/books?id=VQcCV1VuT_cC&pg=PA71&lpg=PA71&dq=mit+atilla+learns+to+walk&source=bl&ots=n9YkssitMh&sig=zYJ-SRu4KZ7IsWXTPAWeXHVMqCY&hl=en&ei=gZxzS-HeCJCI8Aahg4ydBQ&sa=X&oi=book_result

      • Re: (Score:2, Funny)

        by uberchicken (121048)

        Waaaaay to completely ignore how cool this is

      • by DrSkwid (118965)

        You want to make your way in the CS field? Simple. Calculate rough time of amnesia (hell, 10 years is plenty, probably 10 months is plenty), go to the dusty archives, dig out something fun, and go for it. It's worked for many people, and it can work for you.

                        -- Ron Minnich

    • Re: (Score:3, Funny)

      by Gerzel (240421) *

      Not true the AI has not surpassed the abilities of a drunken man as the drunk is doing it with considerably fewer legs, usually 2 to 4, 5 if they are lucky.

      • by PiSkyHi (1049584)
        Hey mod pointers! mod systems just don't work with subtle humour, I guess by definition.
  • given the AI article just a few stories down.

  • by bipbop (1144919) on Wednesday February 10, 2010 @09:48PM (#31094880)

    These six-legged robots can dance! Hexapod: Best of Dance 2009 [youtube.com]

    Yeah, this in no way lessens the accomplishment of a robot actually learning to walk, but I figured it was half on-topic, half cool-as-hell so I'd post it :-)

    • Yeah, this in no way lessens the accomplishment of a robot actually learning to walk, but I figured it was half on-topic, half cool-as-hell so I'd post it :-)

      Yeah I was rather impressed with that myself and was curious to see what he was using for processing, sensors and etc... Apparently it was an Atom, maybe TFA said that but I'm not down with FOX links.

      http://www.engineering.arizona.edu/news/story.php?id=114 [arizona.edu]

      I would of been much more impressed if he would of done this with something akin to AVR and read about it on Society of Robots instead of FOX but that's neither here nor there...

    • by catd77 (1743104)
      I think robotics have made exponential leaps in recent years. Now, Dean Kamens FIRST competition has high-schoolers making advanced robots. Pretty soon, with innovation like this happening every day, I'm sure there will be robotic "Public Workers" soon enough.
  • Stumbling around on six legs isn't very hard. Almost any vaguely reasonable leg movement strategy will work. Look at "Stiquito" [stiquito.com].

    2010 is a little late to be doing a six-legged crawler. They're fun to build, but you don't issue a press release.

    • by iamhassi (659463) on Wednesday February 10, 2010 @10:23PM (#31095152) Journal
      "Stumbling around on six legs isn't very hard. "

      That link didn't show what stiquito could do. Here's a video [youtube.com].

      "2010 is a little late to be doing a six-legged crawler. They're fun to build, but you don't issue a press release."

      I think parent is right, seems six-legged robots have been around forever. An electrical engineer senior shouldn't have a problem building one of these without a kit, although it looks like he might have used this kit [hexapodrobot.com]. Sure the legs look a bit different, but the placement of servos, etc look the exact same, and before someone says "how many different ways can you build a hexapod robot?" there's many different designs [google.com]
      • Re: (Score:3, Interesting)

        by SpinyNorman (33776)

        He built it for a cognitive robotics class, so the emphasis was on the software, not the hardware (it uses a webcam and optical flow calculation for movement detection, for feedback into the learning algorithm). The FOX article is a horrible source for this story, but if you Google a bit you can find that he used a 3-D printer to build his own legs for the slick version shown - definitely not a kit!

    • Isn't this the MSR-H01 Hexapod [hexapodrobot.com]?

      here's the student's video [youtube.com]

      Here's video of the MSR-H01 Hexapod:
      video 1 [youtube.com]
      video 2, at 1:35 it does similar "body wave" movements [youtube.com]


      The legs look different, but the student does say on that youtube description "This is a demonstration of the new leg design which is much more solid than the previous design."
      • Re: (Score:3, Informative)

        by ejtttje (673126)
        No, for one the hexapod in question is made with Dynamixel servos from Robotis, not the "hobby" servos on the MSR... it's also running a dual-core Atom CPU (Z530 if I remember) vs. a PIC microcontroller, and he is actually using that CPU to do some nice vision processing (optic flow). The MSR is a nice hexapod design, but this newer hexapod is quite a bit more powerful.
        • by iamhassi (659463)
          So what you're saying is he built the exact same thing that has existed for many years, but his needs a dual core 1.6ghz processor? I still fail to see what this does that wasn't done before. If I strap a quad core to my coffee maker to run the little digital clock have I made a "better" coffee maker? I would think the "old" design is more impressive because it's more efficient, it can do so much with so little processing power.
          • by ejtttje (673126)
            The old version doesn't have vision processing... vision processing is a whole 'nother ball game of power and complexity. Adding the extra CPU opens new doors for doing interesting research, he's aiming higher than a remote controlled puppet. For instance, I'm doing research in manipulation and motion planning with a similar robot, and quite appreciate his design.

            But yeah, if all *you* want is a coffee pot, then don't bother giving it a brain.
          • by ejtttje (673126)
            Also the Dynamixel servos [robotis.com] are an important upgrade IMHO

            This is because they have digital communication with feedback, so the robot can actually sense its "muscles". The common hobby servos, even the ones advertised as "digital", only have one-way analog communication, so all the appendages are blind... no idea if it has run into something, and no idea how hard it's working. [*] Also, there's no way to adjust controller parameters in the servo, for example to make it soft for interaction.

            But these serv
    • by Nyeerrmm (940927) on Wednesday February 10, 2010 @10:37PM (#31095258)

      This is similar to a stiquito in only the most superficial way -- its a movable machine with 6 legs. In every other way its different.

      Controls: The stiquito has a single (or sometimes two) actuators, that are placed to mechanically, repeatedly cause the same walking motion. This student's robot has 12 actuators, 2 joints on each leg. This makes the robot much more versatile, but also makes the control problem much harder to solve.

      Learning: A stiquito is dumb -- you attach the SMA to the legs, and put a current through to tighten them. It works exactly the same every time, and you have to put it together in just the right way to make it work. This robot is self-learning (or more exactly, learns through reinforcement). The designer simply creates a fairly simple algorithm that has it try motions and see if it gets it to move in the desired direction, and then learns how to do it over time.

      While I think its fair to say anyone with some mechanical aptitude and knowledge of machine learning could put something like this together, its not exactly a simple feat and is certainly impressive for an undergrad. I don't know of any other self-learning six-legged robots (reflecting my ignorance only), but given the capabilities plus the (likely?) low cost its nothing to sneeze at and could have uses in things like disaster operations.

    • Is that it was impressive enough to catch Intel's attention. It isn't as though this guy was going around to all the news agencies saying "Hey! Look! I made a robot!" No, he made a robot that really impressed his professor. News of it somehow got back to Intel, I suspect his professor probably is friends with someone there, and they said "Wow, that is an amazing little robot. This interests us in particular since it uses our processor." Ok well when a major company is interested in something your university

      • by iamhassi (659463)
        if you RTFA Intel marketing is interested because they can use it to show that their atom processors are good for something other than netbooks, they could care less about the legs and how it processes data.
        • Intel (the company) is interested in real robotics. Yes, it may ultimately be a way to sell more processors, but they do serious research into robotics. See this page [intel-research.net] for example, or this page [intel.com]. They have also been heavily involved in image processing (for robotics and other things) for many years, for example with OpenCV [wikipedia.org].

          Who knows what the marketing department is really interested in besides making Intel look cool.
    • The configuration of the robot is not the accomplishment here, its the fact that robot can actually learn to walk on its own. As the article mentions; if a leg is damaged, it an re-learn how to walk without that leg. Most if not all hexapod robots would have a lot of trouble moving at all if one leg, or even two, were damaged. Based on what this thing seems to be capable of doing, that would only only probably mean a few minutes of down time before it started moving again. IT wouldnt have to be manually re-
  • by calmofthestorm (1344385) on Wednesday February 10, 2010 @10:01PM (#31094984)

    Four legs bad.

  • Eh... (Score:2, Informative)

    by Facegarden (967477)

    A similar design might be used to explore unstable environments such as collapsed buildings or rocky landscapes.

    No, it won't.

    I made a hexapod with 3DOF per leg that could walk in any direction "from scratch" by myself, in high school, for fun.

    Adding some foot sensors is the obvious next step, and I've heard a lot about learning algorithms for walking robots being used over the years.

    Honestly, I'm only bitter because I made something cooler in college but never bothered to post it online, so no one saw it aside from my classmates. But, it was a battery-powered 4 legged walking robot that ran a micro ITX windows XP pc

    • Re:Eh... (Score:4, Insightful)

      by JohnFluxx (413620) on Wednesday February 10, 2010 @10:33PM (#31095230)

      For such a long rant, you didn't seem to given any reason why it won't.

      • For such a long rant, you didn't seem to given any reason why it won't.

        Well, it was a rant...

        But i dunno, it's just nothing *new*. This is no more likely to be the next moon rover than any of the other hexapods people have built. You can buy kits to build this kind of thing for ~$500, and I've been told that if you study it, the learning to walk algorithm is pretty simple.

        I know a great source for something exactly like this that I saw 10 years ago, but unfortunately his site went down a few years ago. (tappotec.net)

        And I did admit, I'm partially just jealous that these kinds

        • by JohnFluxx (413620)

          I'm sure that wheeled robots were around for a long long time before one was sent to Mars or the Moon. So just existing for ages without being used doesn't mean much

    • Re: (Score:2, Informative)

      by Anonymous Coward

      First: your story sounds like a lie.

      Second: you didn't get the point, that piece of software involves computer vision and A.I. Creating a simple neural network for reading characters of moving images is something not even 1% of computer "geeks" can really do.

  • I distinctly remember reading an article in Discover magazine about six-legged "insectoid" bots that taught themselves to walk... nearly 15 years ago.

    • Re: (Score:3, Interesting)

      by Meshach (578918)
      The summary does not do the article justice. This is the first line from the actual article:

      Picture a spider-like robot that teaches itself to walk, can adapt when damaged and watches its maker as he moves around the room. That might sound terrifying.

      The exciting thing is that the robot could compensate when part of itself was damaged and get around/over obstacles

      • Re: (Score:3, Interesting)

        by Facegarden (967477)

        The summary does not do the article justice. This is the first line from the actual article:

        Picture a spider-like robot that teaches itself to walk, can adapt when damaged and watches its maker as he moves around the room. That might sound terrifying.

        The exciting thing is that the robot could compensate when part of itself was damaged and get around/over obstacles

        Actually, that's part of the learning algorithms that have been around for a long time. Since it can teach itself to walk, it can re-teach itself with broken appendages.
        -Taylor

        • by srothroc (733160)
          Imagine a swarm of these things communicating via wireless or 3G, sending walking algorithms to each other while traversing difficult terrain.
          • Re: (Score:3, Funny)

            by Facegarden (967477)

            Imagine a swarm of these things communicating via wireless or 3G, sending walking algorithms to each other while traversing difficult terrain.

            I've built enough walking robots to not be too chilled by that vision. They'll just kinda poke along, really slowly, and then their batteries will die.

            Awesome.
            -Taylor

            • by srothroc (733160)
              Chilled? I was thinking more of S&R possibilities or exploration.
              • Chilled? I was thinking more of S&R possibilities or exploration.

                I knew that word was going to be the one to get a response. Chilled, titillated, interested, whatever word you want to use. I used chilled because a couple posts up someone quoted a line from the article saying that it might be "terrifying" to see the robot move.

                But yeah, search and rescue, I know. Its just that while this hexapod is a particularly nice one, its still somewhat basic. Its unlikely to get farther than a robot with tank treads on it, because they just work really well. That's why the military

            • by Shadyman (939863)
              But we'll likely give them solar panels. To stop them, we'd have to block out the sun...
              • by gnud (934243)
                We know that it was us who scorched the sky...
                ... on the upside, that got rid of global warming.
          • Re: (Score:3, Funny)

            by rockNme2349 (1414329)

            Imagine a beowulf cluster of these things communicating via wireless or 3G, sending walking algorithms to each other while traversing difficult terrain.

            You seem to have made an error in your comment. I took the liberty of fixing it.

          • by adonoman (624929)
            Ideally they'd be made up of thousands of uniform pieces that could assemble into various forms as needed. They should be able to build more of those "blocks" as needed, to create copies, or replicas, of themselves. Also, those blocks should be able to communicate with each other via subspace.
        • Re: (Score:3, Informative)

          by Jon Abbott (723)

          That is correct. Mark Tilden has been doing similarly cool walking robots (but mostly in analog!) for years now. Check this out:
          http://www.youtube.com/watch?v=gM4DitOJdyA [youtube.com]

          I remember seeing a video of one of Mark Tilden's robots (or maybe it was one of Rodney Brooks) and he was able to bend a leg back and it would keep walking successfully with the remaining legs. The beautiful part was that there was no microcontroller involved - it was simple analog circuits replicating neuron functions. The class of ro

      • I agree with Verteiron. This is NOT news. MIT has been doing work in this field since I was in high school in the mid to late 80's.

        When I graduated in 89, MIT Robotics Group had already created a six legged robot that could learn to walk through a distributed reward based algorithm. I believe the robot was called attila.

        Then in the early 90's they produced a couple new "generations" of these robots that could follow objects, seek shelter under chairs based on the shade (shadow) provided by the chair or tabl

      • Re: (Score:1, Informative)

        by Qlither (1614211)
        <quote>The summary does not do the article justice. This is the first line from the actual article:<blockquote><div><p>Picture a spider-like robot that teaches itself to walk, can adapt when damaged and watches its maker as he moves around the room. That might sound terrifying.</p></div></blockquote><p>
        The exciting thing is that the robot could compensate when part of itself was damaged and get around/over obstacles</p></quote>

        The whole point of a
    • by Blakey Rat (99501)

      Or the video game Galapagos. That came out in, what, 1996?

      (Yah, it wasn't a physical robot, but it was a virtual one, and it certainly learned to walk on its own given enough time.)

      • by Cochonou (576531)
        I think that's the point. Haven't _most_ of these experimentations first been performed in a virtual environment (using genetic algorithms and such on many generations quickly), then implemented on a real robot for perfection ?
    • I'm sure they didn't do it the way this one does though - not enough compute power back then. This thing only uses a webcam for feedback to learn walking. It does onboard optical flow processing using Intel's OpenCV library to determine if it's (initially random/uncoordinated) leg movements are moving it forward.

  • by Anonymous Coward

    here's the link to UA Engineering's story w/ youTube video:

    http://www.engineering.arizona.edu/news/story.php?id=86 [arizona.edu]

    or, cnet: http://news.cnet.com/8301-17938_105-10450394-1.html [cnet.com]

    mod this into the ground as flamebait, but why in the hell would one want to read about scientific achievement in an article posted on a cable "news" station's web site (read: all of the cable "news" stations are pure crap), let alone the one that serves as a megaphone for those most hostile to scientific achievement. Let's see, do

  • a book released a book in the 1970's where they had a simple chess like game in basic that had a 3 x 3 array. The computer would make a random legal move.. if the computer did not lose after the move.. it saved the move for next time. if it lost after the move it would remove it from memory or more advanced was to block that move if the board had the same set up. so while he is making a robot learn to walk by its self (cool) the logic process in programing is not new at all.. This was based on a math que
  • The project is actually very impressive. The some of the technical details are here [intel.com] including how the learning algorithm was implemented. Reinforcement learning, I knew it!
  • The robot, which is equipped with sensors on each foot, teaches itself to walk

    FWIW, people have been doing this kind of thing in simulation for a long time.

    Also FWIW, in science fiction movies I have trouble with my suspension of disbelief when armies use the kind of "walkers" you usually see. But one with six or more legs could probably work better than track-laying vehicles in extremely rough terrain.

    Probably still not so hot in soft terrain, though.

  • by burningcpu (1234256) on Wednesday February 10, 2010 @11:47PM (#31095802)
    I saw this a few weeks ago when it was emailed to all the students at UA. Here is a video of the guy who made it, and it shows the robot walking around. The video mentions that IBM bought it from him. http://uanews.org/node/29644 [uanews.org]
    • Re: (Score:2, Informative)

      by burningcpu (1234256)
      Whoops, it was actually bought by Intel. Pretty cool that my first informative mod comes from spreading misinformation...
      • Re: (Score:3, Funny)

        by Abstrackt (609015)

        Whoops, it was actually bought by Intel. Pretty cool that my first informative mod comes from spreading misinformation...

        You're well on your way to becoming an editor here. ;)

  • Am I the only one that thinks it looks kinda like a head crab?

  • ...Seems buggy by design.
  • It will teach itself how to grind gold in World of Warcraft.

  • ... but this exact same thing was a demonstration project at the University Open day ... in 1997

  • ... sine to the rescue.
  • okay it were not the simpsons, it was Cornell University, but there was a starfish in a simpsons episode once...

    http://www.youtube.com/watch?v=ehno85yI-sA [youtube.com]

Maternity pay? Now every Tom, Dick and Harry will get pregnant. -- Malcolm Smith

Working...