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


Forgot your password?
Biotech Robotics Science

"Cyber-Roach" Forces Rethink On Animal Movement 41

Lanxon writes "A team of researchers at the Royal Veterinary College in London has built a 'cyber-cockroach' (a cockroach wearing an accelerometer in a tiny backpack) to try and better understand the movements of many-legged animals. They found that unlike bipedal creatures, animals with more than two legs don't adjust their movements when walking over a softer surface." The academic paper is available from the Journal of Experimental Biology. This research will be helpful in finding better ways for multi-legged robots to navigate difficult terrain.
This discussion has been archived. No new comments can be posted.

"Cyber-Roach" Forces Rethink On Animal Movement

Comments Filter:
  • by Gizzmonic ( 412910 ) on Friday May 14, 2010 @05:59PM (#32213768) Homepage Journal

    That robo-roach used to be a human guy, I think he was Czech? Anyway, he woke up one morning and he was a bug. And just when he started getting used to that, they put some cyber-helmet on him and started doing weird experiments on him! Talk about a shitty life!

  • by mangu ( 126918 ) on Friday May 14, 2010 @06:03PM (#32213806)

    Cyber roaches [google.com] have existed for nearly 25 years. Call me again when you have the cyber T-Rex.

  • welcome this priority realignment on migration from our Cyber-Roach Forces.
    • I'm glad I'm not the only one who read that as Cyber-Roach Forces reconsidering their ambulatory strategies to no doubt increase their already terrifyingly lethal capabilities.
      • Re: (Score:1, Insightful)

        by Anonymous Coward

        I didn't consider that interpretation, but it's far more hilarious.

        I read it like so: Cyber-Roach [invention] forces [a] rethink on Animal Movement. Not as being the literal movement of animals, but a reconsideration of some animal-rights activist group.

        Seriously, WTF. Worst headline ever.

    • Re: (Score:3, Informative)

      by Mat'nik ( 1291540 )
      The animal movement really needed this rethink.... i'm glad someone finally forced them. Good job Cyber-Roach!
  • This research will be helpful in finding better ways for multi-legged robots to navigate difficult terrain.

    I thought I saw a video where they took an "AI" Unit, gave it a bunch of simulated legs, and told it to brute force its way into walking to move a distance, knowing only how to move joints. I believe it was a 6 legged thing, sprawled out on the floor to start with. They were kind of hoping for a spider-like walking thing to be procedurally generated this way, but what they found was that it kind of slinked itself along more like how you would imagine a starfish moving along the ocean floor.

    I found it quite

    • by u17 ( 1730558 )
      It's been done many times, not just in simulation but in real robots. See, for example: http://www.youtube.com/watch?v=RZf8fR1SmNY [youtube.com] . Learning basic movement, or learning to navigate in a maze, without knowing what your available effectors do, but being guided by some kind of reward signal, are problems commonly solved by "reinforcement learning" techniques. The "Q-learning" instance demonstrated in the above video may seem brute force at first, as it begins with random exploration, but as it progresses, it
  • A bit of a stretch (Score:5, Insightful)

    by tomhudson ( 43916 ) <barbara.hudson@[ ... m ['bar' in gap]> on Friday May 14, 2010 @06:06PM (#32213840) Journal

    ... animals with more than two legs don't adjust their movements when walking over a softer surface ...

    ... should probably read "some insects don't adjust their movements when walking over a softer surface". To extend this claim from cockroaches to all animals is so stupid it doesn't even rise to the level of "bad science."

    • by spazdor ( 902907 )

      Yeah, I thought this same thing.

      "Trivially disproven using a cat and a blanket."

    • Re: (Score:3, Interesting)

      by Locke2005 ( 849178 )
      Not "more than two legs", animals with 6 legs don't need to adjust their movements because if they only move 3 legs at a time, they are always inherently stable. Animals with 2 legs, and to a lesser extent animals with 3, 4, or 5 legs (depending on gait), do need to adjust their movement for different terrains to avoid falling over. Yes, if you are going to build a "walking" robot for harsh terrain, 6 legs appears to me to be the way to go. One legged robots [mit.edu] -- not so much.
      • by EdZ ( 755139 ) on Friday May 14, 2010 @07:14PM (#32214654)
        Actually, cockroaches can get over some crazy obstacles with little to no change in gait, That big honking joint in the middle? With just that, they can clime obstacles as tall as they are. Yes, they can switch to a wave gait for really big stuff, but most of the time they simply move their muscles in exactly the same way, and rely on the dynamics of their leg joints to conform them to the surface passively. It's an extraordinarily elegant and efficient way of moving. There's a lot of work in the field of passive-dynamic robotics aimed at modelling this sort of movement, allowing robots to move using legs with far less energy than they do at the moment, by designing them so a lot of the work in moving the limbs is done massively.
        Lookup some of Dr. Roger Quinn's work. I can't find the videos he showed at a recent lecture in the UK, but they demonstrated how the much-maligned Whegs are really a lot cleverer than they look.
    • I think even "some insects" is a stretch. It's not scientific even to say "all cockroaches" until they try it out on more than one.

    • Yeah, just one question, though;

      How many land animals with six legs or more that aren't insects are there?

      • Wait, I just remembered the crustaceans....

        Sorry, my bad, carry on.
      • Spiders - 8 legs. They're arachnids, not insects. Ditto for scorpions, ticks, etc.

        Mammals - Siamese dogs, cats, etc. (Ever see a 6-legged chihuahua? Kind of gross, actually. Still-borne, in a jar. It's mother was BIG for a chihuahua - more like a largish cat)

    • Hey, they said "animals", not "all animals". As long as you find a couple of them somewhere, you're good to go.

    • My first thought was maybe things, that are really lightweight, don't make any adjustment when traveling over soft surfaces.

  • Horribly, this research will likely contribute to a successful implementation of the human centipede [wikipedia.org].

  • So why the lame photoshopped picture of a cockroach with a backpack? Is TFA unable to produce any documentation at all from images to data?

    • by Tablizer ( 95088 )

      I noticed the same thing. The roach is blurrier than the pack. What they did is misleading, and poorly done.

  • was getting the roaches to use the tiny hydration packs in the backpack.

  • i thought this story was about e-weed.

  • how this knowledge will be used to improve on this beast:
    http://www.vubx.com/featured/hexapod-ant-robot-lifelike.html [vubx.com]

  • Journalists, you do realize the cockroaches do not actually carry a backpack with an accelerometer in it, right? I am quite certain that the scientists simply glued the accelerometer to the cockroach and called it a day!
  • I thought it was referring to the animal rights movement and why did they care about cockroaches? Forcing them to wear backpacks?

They are called computers simply because computation is the only significant job that has so far been given to them.