Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
Toys

Mathematica and BattleBots 80

Posted by michael
from the math-only-mostly-useless dept.
hesheboy writes "Wolfram.com has a story about building a battlebot with Mathematica: 'October 28, 2002--Looking for action with brains-over-brawn appeal? William McHargue, a freelance physicist and long-time Mathematica user, is one of many who find this combination in BattleBots, the new fighting-robot craze. "With BattleBots, one can be aggressive and yet nobody gets hurt," says McHargue. Recently, McHargue was featured in Mechanical Engineering magazine for work on Tesla's Tornado, his BattleBot.'"
This discussion has been archived. No new comments can be posted.

Mathematica and BattleBots

Comments Filter:
  • While reading (Score:5, Informative)

    by jukal (523582) on Thursday October 31, 2002 @03:21AM (#4571061) Journal
    ...remember that Wolfram.com [wolfram.com] the site on which the story resied == Mathematica. The company whose product [wolfram.com] Mathematica is. So, do not expect to see something unprejudiced. It's an interesting story anyway :)
    • Re:While reading (Score:3, Insightful)

      by giminy (94188)
      Chuckle. Was kind of wondering about that. It'd be like saying "Joe Smith, a long time Microsoft Word user, has just released his latest book."

      People aren't interested in the tools used to make the product, unless they're the company that makes the tools and are making a press release :). TBH I'm surprised that hesheboy's email address isn't something@wolfram.com ;-).
    • Does it really matter what software he uses? I think that the robot itself is the only interesting thing, not which software he uses to model it... It's like me running a story on how I wrote my latest software project in C. wow
  • hmmm (Score:5, Interesting)

    by lingqi (577227) on Thursday October 31, 2002 @03:23AM (#4571062) Journal
    Maybe he is onto something design-wise - but I don't think it's "interesting."

    What I mean is (drawing on real-life examples) that while bacteria and viruses (yes it's spelled viruses, see here [dictionary.com]), I don't really think that's what we are looking for when doing battlebots.

    for the longest time, rambots (bots that basically has a lot of power and a wedge shape) would win consistently. This guy's little contraption is not much different. the bot still depends on a very rudamentary skill to attack / defend. - the only difference is that he usese Mathematica for modelling vs. say, ProE (which I think would be better anyhow).

    real brain over brawn would be, let's say, an (almost) universal manipulator, and enough sensors, reactory circuits, and capability that the robot will make reasonable decisions to duck, block, parry, jump, or just (calculatedly) take an attack, and then be able to exploit the other robot's weakness at the same time.

    • Re:hmmm (Score:3, Interesting)

      by Nephrite (82592)
      I agree mostly but you really describe the consequence of a real problem. The basic problem with the battlebots is that they're just too damn err... strong. Who needs that manipulator you describe if just hacking and bludgeoning is more effective? I would suggest introducing some restrictions on the bots' armor strength so that using brute force would damage the aggressor itself (if you use force too excessively of course), thus promoting use of more sophisticated devices and algorithms in the bot construction.


      As to our 'bot-of-the-day' it is just another hard thing bashing on its opponent. Also, I just don't see anything special in using some math software for designing it. After all most engineers calculate their inventions before building them.

    • by Dasein (6110)
      I don't care about the software he uses. The design is neat. It's the first spinnerbot that I know of that spins the entire chassis instead of just shell.

      To do this the wheels that the bot spins on have to brake at precise intervals to provide the ability to do anything but just sit there and spin. That means he probably has some form of onboard computing.

      BattleBots is neat but one of the things that's always detracted from it in my mind is that the bots always seemed like big, strong, remote controlled cars with no intelligence. This seems like a small step towards intelligence and may actually raise the bar.
  • Hrmm (Score:5, Funny)

    by acehole (174372) on Thursday October 31, 2002 @03:30AM (#4571066) Homepage
    So when can I expect to hear the annoucement of a BattleBot weighing in at 3.141592653589793238462643383279 pounds?

  • by the.jedi (212166) on Thursday October 31, 2002 @03:35AM (#4571072) Homepage
    .....Is to design an evolutionary program that would pick some basic designs (wedge, saw, spin, etc...) and have them do battle several thousand times then use natural selection to mix the properties of the most successful robots and greate a new generation of robots then repeat as many times as possible till you get a robot that is a highly evolved killing machine.

    I don't think this would be incredibly hard to do. They I believe they already had a computer evolve a robot that could walk so now we need to evolve a robot that can Smash.Oh and i'd be coold if it could steal the defeated robot's parts and build onto itself. I suppose that would put it over the weight limitations though.

    On second thought they'd probably just start hunting human beings and that wouldn't be cool at all. Guess I'll just put down the wratchet and the C compiler and goto bed.
    • why simulate the battle, hook it up to some manufacuturing machinery and play it out for real.. far more entertaining
    • .....Is to design an evolutionary program that would pick some basic designs (wedge, saw, spin, etc...) and have them do battle several thousand times then use natural selection to mix the properties of the most successful robots and greate a new generation of robots then repeat as many times as possible till you get a robot that is a highly evolved killing machine. No, the best way to build a battlebot is to spend half your time building and half your time driving.

      Let's face it, nearly every bot on that show could be twice as deadly if they just got someone talented enough to drive it and operate the weapons. It's like putting a $3000 stereo system in a $1000 car...all that work for nothing.
      • No, the best way to build a battlebot is to spend half your time building and half your time driving.

        This is precisely why I refuse to watch battlebots. As long as a human is driving it, it's nothing more than an RC car. A competition of autonomous robots, OTHO, I would be interested to see.
        • Righto to both of you. Thanks. Yeah, I guess most autonomous robots are pretty boring. But still, I think one could make a spastic fighting bot that would work on its own. Especially given that it will be operating within a closed ring on a flat surface, and all it will have to do is shoot anything that moves... :-)
      • Actually, it's like investing $3000 in subs and amps for the stereo in a $1000 car, then forgetting that there's some very important sound in the "over 500 Hz" range (and that a vibrating trunk will overpower even the cleanest bass outside of that POS car, making the passers-by who are supposed to be impressed actually think that you only spent $50 on a crappy buzzy stereo).

        How's that for a run-on sentence?
    • Domon Kashu: I must defeat The Dark Gundam!

      the.jedi: you what? hey! it's just for battle-bots! put that sword away!
    • Dude, this would require a heck of a lot of physics knowledge.

      Its a cool idea, definately. But you have to program out the physics of *every* interchangable component, including the dynamic physics (what happens to objects when they are struck, moving, rolling, etc - even what happens to batteries when subjected to a certain amount of force in a particular direction). It would be an incredibly complex model that would need a lot of computing power.

      Besides, since in BattleBots humans are controlling the robots, you would have to make an AI to act as a human in controlling the robot during the various evolutionary rounds. And once you have an AI that good, you might as well include it in the real robot. :-)

      Cheers,

      Costyn.
  • I'm a physicist. I've used Mathematica. I think it's a great program.... But, I don't like Wolfram's politics. He and CalTech had their battle back in the day about the engine that makes symbolic math possible. Wolfram won, now he's running the big proprietary business.

    Are there any nice Free or Open Source alternatives. I know that maple and matlab do this stuff to some extent, but I don't know their licenses. Seems like there should be a project. That software is expensive as crap.

    • by Anonymous Coward
      No.

      Use the right tool for the right job. Choosing software based on politics is like choosing a hammer for its smell.

      Also, Mathematica does nothing for you that you couldn't do yourself with a pencil and some paper. Convenience costs money. If you can't afford it, you must not really need it.
      • Choosing a hammer for its smell.

        I think that's a bad argument. While it may be a valid argument for the one time project, it is not a valid argument when you are talking about science and building up a vast source of tools that can be used by all. I agree mathematica is nice, and convenient, and I don't have a problem with paying for the convenience. The problem is that it is proprietary software and there are arguments against that in general.

        As for the whoever says I should write my papers in Word ... that's ludicrous. i'd spend the rest of my years clicking buttons in the equation editor, I couldn't put my stuff on xxx.lanl.gov, and I'd be looked at as a fool. Everyone in high energy physics uses unix basically. CERN supports gcc and g77. They wrote paw, it's under a modified GPL... you try and break out SigmaPlot and excel and you'll get laughed out of Geneva.
        • The problem is that it is proprietary software and there are arguments against that in general.

          Yes... but those arguments are universally laughed at by those of us who understand that computers are tools to be used to accomplish productive work, and who believe that people who deliberately choose an inferior tool because it comes with source code deserve what they get.

          The general arguments against proprietary software, most of which were advanced by Stallman, are all really unconvincing.

          Everyone in high energy physics uses unix basically.

          Everyone in medicine and biology uses Windows or Mac, in about a 50/50 (at most 60/40) fraction. If you try and break out Gnuplot and LaTeX in a medical or biological research facility and you'll probably be politely told to use tools that are compatible with what everybody else uses. If you then start talking about politics... well, see the above poster who talked about research assistants being fired for making a nuisance of themselves.
      • Freedom has no price. The right tool for humanity is freedom. You can ignore that fact and freeride from the hard work lots of "irrational idiots" that are sacrificing themselves to provide you freedom.

        Hey, this is not theoretical. The "right tool for the job" today may have a huge impact in what you can do in the future. Look at the .doc, .xls problem. The right tool for the job costs everyone in the world $400 to just be able to use the standard documents, spreadsheets and presentations.

        But your tip is aprecciated, because there will always be people working for your freedom for free.
    • by Anonymous Coward
      If you want something like matlab that is open source (GPL) you can take a look at Octave (www.octave.org). Nothing symbolic in the basic design, but maybe some component that is symbolic and runs in octave? I hav'nt looked into that program that much.
      • The problem with that program is that it plots data out using Gnuplot which quite frankly sucks in comparison to the plotting facilities in the real thing.

        As hard as it may be to comprehend for some people (RMS in particular), sometimes it actually is worth paying for proprietary software. I have bought Intel C++ and Intel Fortran compilers, Matlab and Labview. Why? Because they completely outclass any open source alternatives. I have a job to do and I want to do it in the best possible way and I will buy whatever I have to to accomplish that goal.

    • by krazyninja (447747) on Thursday October 31, 2002 @05:18AM (#4571134)
      You can try Scilab from here [inria.fr]. It is a free scientific computation tool, feel-like-matlab clone.

    • Why don't you try the GNU Free Software Directory? Take a look, there's some math packages here [gnu.org] on it already!
    • I've used all 4 main tools: Mathematica, Maple, Mathcad, and Matlab. All 4 have their various strengths and weaknesses. Overall, if I had to choose one as "the best or most comprehensive", I'd have to choose Mathematica. Now, if only symbolic math is important to you, then try Maple. It's a good product and strong with symbolic math. You can download a free demo from them.
    • I do not know of any OSS 'symbolic' maths program out there, but there is a very competent matlab close called Octave (www.octave.org).

      It uses the good old Fortran kick-ass linalg libraries for counting and gnuplot for the graphics.
    • Maxima [sf.net] is a GPL'd symoblic algebra system.
    • by Anonymous Coward
      Yacas [sourceforge.net] -- A symbolic computation engine similar to Mathematica or Maple. It has a Lisp core, with plenty of syntactic sugar. Released under the GPL.

      Octave [octave.org] -- A damn fine piece of work for numerical computation. IMO, it beats MatLab any day. Released under the GPL.

      Maxima [sourceforge.net] -- a descendant of Macsyma, which all True Math Geeks remember. It's a symbolic computation engine with a Lisp core, like Yacas. Released under the GPL.

      JACAL [mit.edu] -- another symbolic computation engine with a Lisp core. Released under the GPL.

      GAP [anu.edu.au] -- a system for doing abstract algebra and combinatorics. This is really only of interest to a limited subset of mathematicians. However, it is incredibly good at doing what it does. GAP is under its own license, which I'm fairly certain would classify as free to RMS.

      There are many others, but these are the most mature that I've dealt with. If you're looking for a pretty front-end, Maxima has one, there's one for Octave called G-Octave [kstraight.net] (uses Gnome), and there's one for GAP called XGAP [anu.edu.au]. None of them match the purtiness of Mathematica or Maple, though. There is TeXmacs [texmacs.org], a rather impressive TeX-ish WYSIWYG. With some effort, you can make it serve as an input/output mechanism for any CAS. However, I recommend against using it for its intended purpose as, although its rendering is very impressive, it is a big step backwards for structured documents.
    • Damn it is impossible to delete posts in slashdot. I meant this link [slashdot.org]
  • Website (Score:4, Informative)

    by Kj0n (245572) on Thursday October 31, 2002 @04:59AM (#4571117)
    Found on Google [google.com]: the official website [mchargue.org].
  • by krazyninja (447747) on Thursday October 31, 2002 @05:28AM (#4571139)
    ...BattleBots safety regulations required him to perform an analysis to prove that the laser would not harm anyone viewing the fight. McHargue performed the calculations for this analysis and typeset the report using Mathematica....
    If the rules are so strict, this raises a legal question for most mathematical software. Consider this scenario: Due to a bug (which could have been accidental), mathematica reports an "unsafe" value to a "safe" value.
    2. McHargue uses this unsafe laser in his bot.
    3. Somebody gets hurt by viewing his fight.

    Legally who is responsible? Wolfram? McHargue? The organisers? What???

    • > Please do not look into laser with remaining good eye.

      ~some things just have to be said.... this just might not be one of them.
    • You can be sure that the software license for Mathematica contains a limitation of liability clause, indicating that the results should NOT be implicitly trusted for calculations where saftey is a concern. It's boilerplate on pretty much any software license.

      I would also assume that the battlebots organizers would be the ones held liable, which is why they requested the documentation in the first place. If someone sues them for damaged eyes because of the laser use, they have this document to show that they had every reason to believe it was safe, and they were not negligent in assuring audience saftey.
  • "In the BattleBox, Tesla's Tornado is a 117.9-pound block of spinning, smashing steel"

    Anyone know how much Katz weighs? If it is more than 117.9 pounds, just imagine the possibilities!

    It shouldn't be too hard to retrofit him with the appropriate wheels and circuitry. Just imagine the possibilities of a spinning, smashing Jon Katz!

    (woah...and what if we made a Beowulf cluster of 'em)
  • Is that once the robot's unconquerable might is proven in Mathematica, it becomes unnecessary to actually build the device.

    "An excercise for the student", I believe they call it. ;)
  • I recently read Wolfram's book, and was most frustrated by the way that it is tied so closely to Mathematica. Mathematica is a very impressive, very important analysis tool, and is REALLY FSCKING EXPENSIVE.

    Oh, by the way it was the New Kind of Science book, not the Mathematica book that I read ;)

    At any rate, I found some cool analysis tools that people should check out as alternatives to Mathematica for analysis and visualization of everything from battlebots to cellular automata. Without further ado:

    1. PDL [perl.org]
    2. R [r-project.org]
    3. PGPlot [caltech.edu]
    4. GRASS [ibiblio.org]
    (just to name a few)

    PDL is the most directly analagous to Mathematica or Matlab. R is, of course, like S/S+. PGPlot is for visualization. Grass is mostly for geostatistics/GIS. But it's cool enough to throw in the mix.

    Anyhow, hope this helps someone out. Go forth and make a battlebot.

  • If you're looking for a BattleBots for the masses, head over to Botbattle [botbattle.com] and try it. You program a bot in a basic-like language and watch it fight. It's kind of cool.
  • Click the link in my signature if you're interseted in a free program that lets you design and play with virtual radio-controlled toys. I hesitate to use the word 'robot' as the control system language isn't expressive enough for much more than joystick control of walking critters, but it's kinda fun.

    An open-source multiuser "arena" runtime is only a matter of time.

  • "I love Saturday morning cartoons, what classic humour! This is what
    entertainment is all about ... Idiots, explosives and falling anvils."
    -- Calvin and Hobbes, Bill Watterson

    - this post brought to you by the Automated Last Post Generator...

It was kinda like stuffing the wrong card in a computer, when you're stickin' those artificial stimulants in your arm. -- Dion, noted computer scientist

Working...