Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Robotics Technology

OmniTread: A serpentine robot 150

Posted by samzenpus from the crawl-anywhere dept.
karvind writes "Physorg is running a story about OmniTread: a serpentine robot designed to traverse extremely difficult terrain, such as the rubble of a collapsed building. The 26-pound robot is developed at the University of Michigan U-M College of Engineering. It moves by rolling, log-style, or by lifting its head or tail, inchworm-like, and muscling itself forward. Link to videos. Check out there other robots as well."
This discussion has been archived. No new comments can be posted.

OmniTread: A serpentine robot More

OmniTread: A serpentine robot

Comments Filter:

  • Crikey! (Score:1, Offtopic)

    by Dancin_Santa ( 265275 )
    Someone call Steve Irwin!

  • Clarification of how this works (Score:5, Funny)

    by MobyDisk ( 75490 ) on Wednesday March 23, 2005 @11:37PM (#12032534) Homepage
    From the article:
    These treads prevent the snakebot from stalling or becoming stuck on rough terrain because, similar to a tire touching a road, t the treads propel the robot forward like a tire touching a road.
    Well, that sure made it clear!

  • cool (Score:3, Interesting)

    by dteichman2 ( 841599 ) on Wednesday March 23, 2005 @11:38PM (#12032540) Homepage
    I imagine that with more development, this could lead to subways/trains without tracks. Or, perhaps "smart" cars that "know" how to handle obstacles and avoid collisions.

    • Re:cool (Score:4, Funny)

      by Jeremi ( 14640 ) on Thursday March 24, 2005 @01:40AM (#12033183) Homepage
      Or, perhaps "smart" cars that "know" how to handle obstacles and avoid collisions.


      Great, just what I need... SUVs crawling over my house at night.

    • Perhaps I'm a little slow, but why did the parent get modded as a troll? Just curious.
    • The passenger cars are deliberately made somehow hard to drive. While the automatic system cannot handle every possible situation, the driver has to be kept alert.
    • Or, perhaps "smart" cars that "know" how to handle obstacles and avoid collisions.

      Great! So now we'll get SUV who'll gladly climb over any VW bugs unfortunate enough to be in its path. The SUV will handle the situation just fine, but the VW bug will be slightly flatter than before!

      • That's not true. According to the commercials, the dome design will improve structural integrity and prevent worm-robot-SUV-flattenings. ;-)
  • Vince Ricardo: Serpentine Shelly. Serpentine!

    From "The In-Laws," possibly the best Peter Falk movie ever, right next to "The Great Race."

  • Spheres with tentacles are better (Score:5, Interesting)

    by GAATTC ( 870216 ) on Wednesday March 23, 2005 @11:39PM (#12032544)
    From a comment posted below the article: "Arthur C. Clarke had it right --- spheres with tentacles; _that_ is the ultimate in agility and mobility, for robotic design. Plus, such units can easily link together to form a much greater whole, if required --- they could perform nearly *any* engineering, construction, or transportation task."
  • Am I really that unobservant or is the hardware section [slashdot.org] new?
  • You're just trying to set a world record for /.'ing aren't you?

    Can I suggest to the editors that if you know you are linking to a video, you simply place a link to a mirror there as well.

    Having said that, the site is holding up remarkably well - I've still got 14kb/s..... uh oh 13..... 12..... damn.

    • Doesn't mirrordot automatically mirror things from slashdot? Wouldn't that make some kind of infinite loop? Mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors of mirrors... [no carrier]
    • I think this was the same UMich server that got taken out by a /. story around exam time last December. It's probably close to midterm/exam time there now. I smell a conspiracy... "but the server was slashdotted! How could I look up the course materials?"
      • no, no

        that was eecs.umich.edu. eecs is probably some dual pII in a closet somewhere. It serves up static pages, mostly for internal usage.

        engin.umich.edu is a bunch of heavy iron suns. Probably right now there are more than a few instances of matlab and some simv procs running
        • Guys... Slashdot wouldn't taken down umich... please let me know when this supposedly happened... I'm an engineer on campus and I've seen some of the servers... I use them every day for all my homework... engin.umich.edu and the various subservers are continuously connected to my desktop with ssh for the AFS serverspace as well as for compiling/running/writing my code for my Programming classes. Around exam time someone fucked up with the news-servers... they were switching them around and brought themselve
          • Um, why you responded to me instead of the parent is sort of confusing. when the London Law game got posted on slashdot, it destroyed EECS.umich.edu. I was in the 4th floor eecs opteron lab at the tme, and we were all rather pissed. (I was in 470 and needed some of the specs on the server) I don't have an eecs account, so all I actually know about it is that it is running solaris 8.

            I've been going through the archives, and I found a dupe of the story, but not the actual one (unless comments got ripped
          • PeopleSoft supplied us with crappy server intensive webapplications for our class selection interface


            By god, my school isin't the only one then, I could not for the life of me figure out why my school bought such shitty software from them. Apparently my school wasn't the only one scammed by their PR.

            Peoplesoft = worse web apps I have ever had the displeasure to use.
            • The sad thing is that we kindof knew that... but they had an agreement where they got to update our software at will... we had already had a working version of software but they decided the needed to update ours. With a completely new I'm only assuming barely tested version, because it completely fried for awhile, and even released private data on 2 or 3 occasions. Privacy breech due to bad code... and the funny thing is that I heard that the company just recently got sent on a move to dissolve (Adobe, or O
    • I'm still downloading it at 95KB/s...
    • 700Kb/s here.. impressing..

  • I wonder (Score:2, Insightful)

    by Yeshua ( 93307 )
    How modular the design is.

    It is obviously made of 5 (reasonably) identical parts, but I wonder if you can (in theory) make a robot of this type as longs as you want just by `tacking' on a new section (of course this ignores drive train problems).
    • it shouldn't be too hard to figure out a way to make the drive train adaptable to additional sections, look at some laser printers (ie. IBM InfoPrint 1120) you can add an additional tray to the device by just sticking it at the bottom of the stack, you don't even have to worry about plugging anything in. You would just have to be mildly creative about it.

    • Re:I wonder (Score:4, Insightful)

      by AeroIllini ( 726211 ) <aeroilliniNO@SPAMgmail.com> on Thursday March 24, 2005 @12:13AM (#12032740)
      It is obviously made of 5 (reasonably) identical parts, but I wonder if you can (in theory) make a robot of this type as longs as you want just by `tacking' on a new section (of course this ignores drive train problems).

      It actually wouldn't be too hard to give each section its own motor and power supply. The connection between each section would include an umbilical capable of transmitting data, and perhaps even extra voltage if one of the power units (batteries) was not providing sufficient power. At that point, operation of the robot becomes a software issue; how do you program each pod with identical software that can all work together as a cohesive whole, in real time? If all of the individual components are identical but work flawlessly together, at almost any length, then theoretically the robot could add more sections or shed damaged ones without compromising functionality. Specialized sections could also be attached to the front and/or rear, for digging, drilling, collecting, clearing, bomb dismantling, carrying, or anything else you can think of.

      I noticed in the video that the crew was using a remote to control the robot. (I also noticed the distinct lack of any purposeful log-rolling, contrary to what the editor said--but I digress.) Ideally, of course, such remote operation would not be needed. Crews using the robots--rescue teams, bomb squads, recovery missions--should be able to just tell the robot, "hey, go explore over there. We'll be watching on the monitors, and you tell us if you find anything." The locomotion of the robot would then be completely automated, based on a set of instructions (one of a library of scripts, maybe; use the "find and bring out bomb" script, or the "search for human survivors" script), with the operator just gathering data, and perhaps taking manual control if needed.
      • I interned at Xerox PARC two summers ago and saw a snake-like robot very similar to this [parc.com]. The motion is different---undulating like a caterpillar instead of using treads. But otherwise strikingly similar.

        theoretically the robot could add more sections or shed damaged ones without compromising functionality
        This is how PARC's modular robots worked, which is the really cool part. The modules could be snapped apart and then reassembled into a different-shaped robot. Some models could even reconfigure the [parc.com]

        • Hmm... two summers ago. I worked in the modular robotics lab for a few years a bit before that.

          On-board power is definitely not easy, both because it takes a lot of power to move that much robot around and because batteries make the robot heavier (and larger), which makes the problem even worse.

          I'm assuming your housemate was working on the plastic polybot modules, which were run off of an offboard processing board, or one of the cubic or rhombic models. The G2 hardware had a CAN bus and onboard processin
      • Yeah, totally!

        They showed a prototype of this in that documentary "The Core." They were able to travel to the center of the Earth and restart the core's spin, thus saving Earth!

        Eat your heart out, Jules Verne!

    • Re:I wonder (Score:2, Informative)

      by Nebu ( 566313 )

      How modular the design is. It is obviously made of 5 (reasonably) identical parts, but I wonder if you can (in theory) make a robot of this type as longs as you want just by `tacking' on a new section

      If you go into the slideshow on the site, there's an image [umich.edu] of the robot composed of 7 sections instead of 5.

  • Ooh, snake! A snake! (Score:5, Funny)

    by FlyByPC ( 841016 ) on Wednesday March 23, 2005 @11:44PM (#12032582) Homepage
    Badgerbadgerbadger... (C'mon -- *someone* had to say it...)

  • It's so simple. (Score:3, Informative)

    by Anonymous Coward on Wednesday March 23, 2005 @11:49PM (#12032611)
    Their other robots are there. They're amazing.
  • ...but I'm sure we will soon. Everyone should check out this rad 7min video [umich.edu] of this thing in action. Very cool. =)

  • Memory Shape Alloys (Score:4, Interesting)

    by TimeTraveler1884 ( 832874 ) on Wednesday March 23, 2005 @11:54PM (#12032637)
    I have always wondered why robotics engineering has not taken more approaches using memory-shape alloys. Mondo-tronics has a product called Muscle Wire [musclewires.com] that has carbon or graphite embedded in the alloy so that it heats up when an electric current is applied. This causes the Nitinol [nitinol.com] to contract as the alloy returns to its "memory" shape.

    Really the only thing I have seen using this form of memory-shape alloys is just for hobbyist projects, nothing serious. Granted there are some problems to overcome, such as duty cycles and heat dissipation. But most of these could be solved, I have looked into them. On larger scale projects the cost could be prohibitive though.

    The value as I see memory-shape alloys over motors, is that it is almost a solid-state actuator. There really is no moving parts that can wear, other than the alloy itself. And these memory-shape alloys have a very high force/weight ratio - thus making the bulk of most robotics not a function of locomotion.

    • one problem is that Nitinol really doesn't contract that much, usually only 5-7% (if I'm remembering correctly) however one especially clever nitinol crawler was able to move something like 20% of it's total body length per step
    • They are slower, and although lighter, are less power efficient than ordinary motors. So unless weight is the only concern (and battery weight is't a problem), they aren't useful for much on reasonably sized robots. And as soon as your robot gets reasonably sized and isn't on a cord, battery size and weight is *the* problem. Also IIRC nitinol yield stress isn't that great (i.e. expensive *and* breaks when your actuator gets yanked on).

      People are still working on artificial muscles though, and hopefully
    • Perhaps because they're not yet strong enough to beat even a teenaged girl at arm wrestling? [slashdot.org]

  • ed2k link of the movie (Score:2, Informative)

    by Anonymous Coward

    remember to remove the spaces

    ed2k://|file|OmniTread.SwRI-7min.320x240 x30.wmv|39219440|00C932FF9AD4D798E92C05D9869EE323| /

  • serpentine movement suggestion (Score:3, Insightful)

    by thomasa ( 17495 ) on Thursday March 24, 2005 @12:01AM (#12032679)

    Only the treads that are touching the ground should
    move. The others moving in air are wasted motion. That does not seem efficient.
  • "Physorg is running a story about OmniTread: a serpentine robot designed to traverse extremely difficult terrain, such as the rubble of a collapsed building."

    I thought Remote control mice [nationalgeographic.com] were taking that job.

  • Obligatory bash.org quote (Score:5, Funny)

    by dracken ( 453199 ) on Thursday March 24, 2005 @12:13AM (#12032743) Homepage
    what does your robot do, sam
    it collects data about the surrounding environment, then discards it and drives into walls

    -Dracken
  • They need to send this baby over the to the disaster area that is my room. First mission: Pick up my underwear and clothes that need washing. It'll probably break down under that kind of stress though.
  • Oh please, there is no robot design possible with a simple singular traction system that is virtually unstoppable. There are many many area where such a robot with a snake-like tread structure will simple be unable to get anywhere at all. I would be impressed if there was a use for anything but reconnaisance from a robot that can deal with as many situations as this one supposedly can however.

  • Slick. Wonder if.. (Score:5, Funny)

    by bunhed ( 208100 ) on Thursday March 24, 2005 @12:27AM (#12032816)
    you can program it in python?
    • A similar device was developed at NASA Ames. At least for a time part of its routines for getting around objects was written in perl, so I suppose Python wouldn't be out of the question. :)

      http://www.nasa.gov/centers/ames/news/releases/2 00 0/00images/snakebot/snakebot.html
  • Somebody fill out a requisition for the Batterylife Activator!

  • More like a caterpillar (Score:5, Informative)

    by SnowZero ( 92219 ) on Thursday March 24, 2005 @12:31AM (#12032834)
    This is more like a caterpillar than a snake. It uses tracks all over its body (like many small feet), rather than a serpentine motion to propel itself. Props though, as this seems far more practical than robots that actually try to move like snakes or inchworms.

    I've sat through many talks about modular robots that are supposed to be able to do everything, yet rarely do anything well at all (I come from a lab doing what I guess you'd call "specialized monolithic" robots). I think this robot is just specialized enough to be useful (using its treads). The walking snake like robots are normally agonizingly slow, but this robot moves at a reasonable speed for the type of applications you'd need it for. Also, tracks should scale up in speed reasonably well if needed.
    • The thing about robotic snakes is that they can get anywhere a real snake can. My professor's snake-bots can climb straight up a drainpipe, tube, or crack. The caterpillar one only managed a 22 degree angle. Doesn't make "real" snakebots more useful in general, but speed isn't everything, and CMU's snake-bots can really get around...
      • In that case progress is being made and that's good to hear. Perhaps I came off a bit too negative in the GP. Pure modular/reconfigurable robots strike me as fantastically difficult, especially when you start talking about onboard power; Snake robots do have a future as you point out, but they still need some work to be useful for things like urban search and rescue. But then again, what robot doesn't need more work :) At any rate my main point is that we should distinguish caterpillars-like vs snake-li
  • When I saw this, the first thing I thought was how nice a large version of this could be. It seems like it would be suitable for something like the DARPA Grand Challenge. http://www.darpagrandchallenge.com/ [darpagrandchallenge.com]

  • What about controls? (Score:4, Interesting)

    by Statecraftsman ( 718862 ) on Thursday March 24, 2005 @12:43AM (#12032893)
    That thing looks like you need more than a joystick. It'd be really great if they showed how many buttons they have on their control board. By my count the 5 segment version has 9 degrees of freedom. 1 for the front back and the other 8 for the 2 degrees of freedom for each joint.

    Also, does it know which way is up and readjust or do you have to figure that out after it rolls?

  • ... then the cheeseball techno (read: different than non-cheeseball techno) was just too much :-)

    Regards,

    John

  • http://www.scrubbles.net/mead/mead12.html
    I came across this a few days ago, seemed vaguely related conceptually.

  • Snakebots are very fragile! (Score:5, Interesting)

    by Quantum Jim ( 610382 ) <jfcst24NO@SPAMyahoo.com> on Thursday March 24, 2005 @01:10AM (#12033021) Homepage Journal
    My friend Jer Romeiko builds these kinds of robots for a living [pitt.edu]. You can download some cool videos of snakebot action [cmu.edu] at his employer's web site (CMU).

    Snakebots are very fragile. Many times a section would break after a few hours demostration. Jer was working on making each section more modular and easier to build. Apparently the main goal of snakebots for many research labs are for providing demostrations (read: grantbots) and giving new grad students something to do. ;-)

    • I've put together more of those segments than I can count (note that this is Mark Yim's design, from PARC, borrowed by CMU). They're using a /very/ old hardware design, using weak motors and weak parts. (The motors + the weight of the segments certainly have enough leverage to break the frames, but that's another issue.) The modular robotics group at PARC had much, much better hardware years ago, and were working on G3 when I left.

      A short overview of polybot hardware:
      http://www2.parc.com/spl/projects/modr o

  • This is reminiscent of... (Score:1)

    by Anonymous Coward
    A one-time BattleBot that was run in Vegas by one of the guys at ILM. It was designed like a snake and was very capable of moving around just like one (with a nice little "rattler" to boot). While it could move around, the problem with the design is that there is little you can do with it. I'm curious as to how utilities and other devices could be attached to make this useful. I suppose small cameras and lights could be implemented, making it somewhat useful for rescue missions.
  • Check out Xerox PARC's PolyBot [parc.com]; each segment contains its own motor and PowerPC processor. This was on Slashdot some time ago [slashdot.org].
    • See also my recent post [slashdot.org] higher up on this thread. I got to see these things in action as an intern at PARC and they're pretty cool. They don't require manual remote control like the UMichigan versions seem to. And they can be reconfigured (sometimes autonomously) into other shapes as well.

      And from the looks of the CMU site [cmu.edu] linked to above, their project is heavily based on this PARC work.

  • In the whole video, this only happened twice... and each time it looked accidental. The primary means of propulsion are the tracks (movement along its length), and bellows (which allow it to lift its "head" to climb over obstacles).

  • Almost everything they learned making the robot has been patented. So do the students who worked to create this robot end up paying the college to patent the research they performed? Do College's have patent contracts boilerplate on student applications now?
    • Yes, they do. Even when I applied to college in 1991, there was a waiver of rights to intellectual property built into the application. Anything I created belonged to the school, regardless of whether it was for class or on my own time. They even handed out intellectual property disclosure forms in the event we had an idea.. This IP agreement was written so poorly that it even claimed IP ownership to things I created while employed off-campus.

      It is just urban legend, but I have heard that some schools make
  • Anybody know how to explain the photo a the 4th page of the presentation [umich.edu]? (robot having its head and middle segment resting on top "stair", while tail is hanging "in the air").

    Such position looks as if it would only be possible if "head" is substantially heavyer than "tail" (or else it would tip backwards). However, the doc states that the middle segment is the heavyest. Or does it also have the means of pumping liquid for one end segment to the other to achieve more optimal weight distribution? In any cas

    • If you look closely and count the segments you should see that the middle segment is on the edge, since that is hooked and assuming the rest of the snake is equally weighted, on its own you should be able to balance it at that point... I don't really see why thay would only work if the head were heavier than the tail...
    • The center of weight is in the middle, the middle of the robot is extended a couple of inches beyond the 'stair'. So the robot should be unbalanced, right?

      But since the two 'head' segments are nearly horizontal, and the two 'tail' segments are diagonal, the 'head' segments extend their mass further away from the tilt point, and thus shift the center of gravity to the 'head' of the robot.

      HTH
      --Blerik

  • Cool, but... (Score:2, Informative)

    by AliasMoze ( 623272 )
    it's hard not to be a little disappointed with the state of robot technology. We've landed a man on the moon, split the atom, and decoded our genes, yet we're supposed to be impressed by a (human-controlled) robot that can crawl like a caterpillar.

    Don't get me wrong. I realize this is a step forward, but the current state of robots seems so behind others.
    • yea, and I'm going to tycho city on my next vacation..

      Landed on the moon, 26 freaking years ago..

      US can't even get a man into high orbit right now...

      What's behind>?

  • News? (Score:1)

    by Anonymous Coward
    I couldn't access the article now, so I can only base this on the comments and the post, but... is this realy new?
    I mean: http://www.snakerobots.com/main.htm

    or am missing out on something from TFA? =)
  • The robotic snake offered a woman working on the project a robotic apple...

  • Important question (Score:3, Funny)

    by arhar ( 773548 ) on Thursday March 24, 2005 @08:45AM (#12034777)
    Does it grow longer the more dots it eats?
  • Over 100 comments so far and I, for one, am the first to welcome our new serpentine robot overlords?

Slashdot Top Deals

Every nonzero finite dimensional inner product space has an orthonormal basis. It makes sense, when you don't think about it.

Close