Slashdot Mirror
BYO Battlebot
An anonymous submitter sent in: "With the new season of BattleBots coming up (filmed in SF over Memorial Day weekend) everyone is sick of the same-old design of a glorified R/C car. This site has the full design for a bot that runs on an onboard 486 and is controlled off a laptop with quake-style controls! Build your own for around $100 by using mostly old parts." On the other hand, Coolrobots.com has info on how to build an expensive battlebot.
I'm surprised no one has mentioned that Slashdot has their own battle bot in the works. I'm also surprised that I came across this on sourceforge instead of reading it on Slashdot.
--
Why can't I moderate something "Wrong" or at least "Grossly Misinformed"?
From the site:
Yep! This is a 486DX2/66 running Windows 98 on a 340 meg laptop hard drive.
...gives new meaning to "Blue Screen Of Death", don't it?
"Depression is merely anger without enthusiasm." - Anonymous
I say why not use REAL weapons: guns, flamethrowers, EMP devices, etc. That would pique my interest. Have the announcers discuss why a particular bot has to have a Faraday cage around it when the other robot does something. Also, it would bring testing to a whole new level. Fireproofed, shock proofed, etc.
Just my $.02
JoeLinux
All things are possible, except skiing thru a revolving door.
PVC is a bit weak for BattleBots. I used it on my first robot, The Tunabomber, but that was for DragonCon's Robot Battles, where they don't have killsaws or robots like Whacker and Ziggo. Incidentally, my website has a tutorial similar to the one referenced in this story, but with more detail.
An onboard PC is certainly overkill for control. I do give it points for hack value, though. Competitors who want computer control functions usually use the IFI system. This also allows you to use PC joysticks to control your robot remotely (a joystick setup was mentioned in an earlier post, this is almost certainly what was used).
I wonder why the guys who built this robot didn't compete with it. After going through all that effort, it should be worth it to get to the competition, if only to see your robot ripped to shreds.
Finally, please moderate down all those people who talk about how easy it is to build a winning robot, unless they've actually done it. Slashdotters: as with Open Source, it's put up or shut up.
Your design to a real part online: Big Blue Saw
I went to the Phoenix show (1996?), and recently (October 2001) helped Mark Pauline and David Therrien set up for a demo here in Phoenix (at X.E.R.O/ChemLab) to show off one of the pulsejets that was to be used on their new hovercraft (which was supposed to make an appearance at the Ultraviolence show in Phoenix - but it got cancelled thanks to a lovely interaction between the SF Fire Marshal and the Phoenix Fire Marshal - thanks a lot, guys! Bastards!), and also a video showing of various prior shows.
The demo went off without a hitch - though every time I saw the pulsejet running - glowing white hot and screaming like a banshee on speed, being manipulated by Mark and company using asbetos gloves, next to the 50 gallon tank of propane that powered it - I wondered if we weren't all going to end up crispy critters.
If you haven't seen an SRL show - you don't have any idea what you are missing - picture being in the middle of a war zone, along with a crash-up derby, and a lot of fire, heat, and smoke - plus a ton of noise that manages to rattle every tooth and bone in your body, while deafening you despite wearing ear protection, incidentally making your ears ring for hours after the show - and you might have some idea about what happens during one.
I put in that time to help on the show - all volunteer, mind you - hoping to help out later for the real show - and then it doesn't happen! Gah!
Anyhow - yeah - SRL makes Battlebots look weak, weak, weak - of course, SRL does beaucoup planning to keep accidents, etc from happening - while Battlebots is more "anachistic" in that fashion, in a way. I tend to think of Battlebots as a tamed down SRL real battle (whereas the destruction of SRL's machines, while real - doesn't tend to utterly destroy them, as sometimes happens with Battlebots), and a different form of entertainment (plus, without all the smoke, flames, heat, and falling ash - it is easier to see what is happening)...
Worldcom - Generation Duh!
Reason is the Path to God - Anon
I'll be first to admit comedy central's battlebots is dull. My friend came up with a way to spice it up. His solution... your 200lb, saw covered robot vs a sack of fluffy animals (puppies, kitties, or duckies would all do).
I didn't say it was a good idea
For a car, two channels is the norm and you can get a third channel for more money -- but that's it. For more channels, you'll have to get plane equipment or spend a *lot* of money on specialised hardware.
`All digital' isn't required (I assume you mean PCM?) FM or even AM ought to be fine (but BattleBots prohibits AM, so it's moot) -- after all, modern R/C equipment has an effective range of around 1.5 miles -- far further than you can even *see* your plane. For a bot, it's unlikely to ever go more than 100 yards from you. At that range, interference isn't much of a problem, even for AM, unless somebody is on your exact frequency.
One thing to note that they don't seem to tell you -- in the US, airplane radios use the 72mhz band, which the FCC has designated for aircraft only. To use it for a ground craft is *illegal*. For ground craft, you're supposed to use the 75mhz band. (There's also the 27mhz band, but few people use it because it's also used by CB radios and there's only 6 channels there anyways. And there's also the 50 and 53mhz bands, but you need to have a ham radio license to use these.)
Futaba will convert some of their higher end radios from the 72mhz band to 75mhz band for $40.
If you do actually make your own robot, please don't use 72mhz equipment! There may be a flying field a half mile away that you don't know about, and you could crash somebody's plane without even knowing it.
(I emailed the coolrobots.com guy about this, and his email bounced -- mailbox full. Guess it got /.ed ...)
--
Also, you'll be very hard pressed to find a pistol grip controller with more than 3 channels. Of course, if you have a seperate driver and gunner, giving the driver a pistol grip controller and the gunner a standard two stick airplane controller (on two different channels, of course) would probably work great.
Also, the two stick controllers aren't bad for R/C cars at all. I've got both, and while I do prefer the pistol grip, it's not that big of a deal. But then again, maybe I'm biased because I mostly fly R/C planes rather than drive R/C cars.
--
They allow autonomous bots, just none of the autonomous robots choose to compete against the human controlled robots. It's my understanding that there are certain exhibitions fights (such as Mark Setrakian's Mechadon, Snake, etc.) for robots that fall outside the rules but are still entertaining for the audience, so autonomous bots tend to fight there, but I believe they're few and far between. :).
I'm into autonomous robots myself, and am considering getting into BattleBots, but at least with some fly-by-wire intelligence on board, perhaps with some sensors as well.
I think your assertion about 'the current state of AI' is a little misguided. When was the last time you sat down with a compiler and tried to write something to make a robot do anything, much less be intelligent, or engage targets with deadly force at its own discretion? Don't you think there's a reason the military still mostly has 'dumb' weaponry..aside from being guided by GPS and recognizing its target optically in a highly preprogrammed manner, it's still dumb.
Another thing is..the more intelligent a robot is, the less likely you are to want to see it destroyed. What would it say about humans if we took intelligent, thinking beings and threw them in a pit to fight to the death just because they were machines? I think it's better left as an extension of the phallus (not that Battlebots isn't fun
It looks like they put in a computer, then used it to emulate a really dumb R/C controller. Why?
Watching those things, it's clear what's wrong. People are driving them visually, which means too long a reaction time. They need some onboard smarts.
The idea here is to get inside the other guy's OODA cycle , so he's reacting to what you were doing, not what you are doing.
The first step is to get some onboard heading control. Put in some cheap rate gyros, then control orientation with a knob attached to an encoder. Want a 90 degree turn, spin the knob 90 degrees. That way, you'll get the desired heading on the first try.
Second, some kind of system that senses the opponent at close range and maintains position relative to them so the weapons can work would be a big win. Ancient though they are, the old Polaroid sonars would probably work. Use the piezo ones, not the electret ones; they're more rugged. And use separate send and receive sensors, so there's no minimum range. The idea is to make the weapons stay on a target long enough to have an effect.
Because they're all banned. There are reasons for all of them, if you think about it, mostly for the audience's safety and judging purposes. I don't think they're trying to make the bots wussy.
--
We (the /. crowd) can rip off the PVC ghetto bot, cover it in paper machet armor (not like it would make a difference anyway against ramming or flipping), write JonKatz on two sides, write Alex Chiu on the other two, and stick a picture of the AYB alien on the top. Battlebot fodder that represents the three most hated presences on the Internet, getting mauled in its first match.
We'd have to make sure there'd be a visible Windows logo too. Maybe have it light up when the bot dies. I'm thinking BatSignal here.
What would really be great would be if it actually won a match.
Although actually, knowing how junkyards come into play, there IS a decent weapon available - the phallic sonic ramrod. Take any working, but junked vibrator (just be sure to wear gloves...), overclock it (double the voltage, stick a Peltier in there, etc. etc.) attach a sharpened roof nail to the end, and voila - a ramrod that vibrates the opposition apart while offering a superb visual effect for prime-time TV. Extra points for getting somebody to use it as originally intended without serious injury.
Quoth the article: Most battlebots are just glorified RC cars. They use off-the-shelf speed controllers, and stock RC controllers. Unfortunately, this makes them impossible to drive.
It's not the use of R/C car parts that is the problem. A good R/C car is a dream to drive, very controllable. The problem seems to be with the operator's choice of controllers. I keep seeing these sort of twin-stick controllers in use for simple wedge bots with no extra weapons; why the hell don't the operators use pistol-grip style controllers?
I could never really get the hang of twin sticks, but I can pilot a car pretty well with a pistol grip. AFAIK all serious R/C car guys use them, and for good reason.
The Battlebots arena is PLAGUED with really awful driving. I'm sure a heavy bot isn't as easy to drive as an R/C race car, but c'mon, most of these guys could do a lot better, and I think better controls would help. (I saw one guy using a freaking joystick... I think he got his clock cleaned, too. Use the right tool for the job, Chester!) The videos they show of the designers tearing up junk in parking lots... not a great way to practice. A parking lot is a lot bigger than the arena, and poor control won't be punished as much. Especially when you are wrecking a TV or an aquarium, instead of another bot.
I know some bots need more than just movement controls, and a 2-channel pistol grip isn't adequate. Nonetheless, it sure looks like some teams are sabotaging themselves with a poor choice for mobility controls. Mobility is life; precise driving should be the first requirement for any bot.
Seems to me a team should have one driver and one gunner (yep, some do, I know), or perhaps one operator using some innovative controls like footswitches to operate the weapons. A pistol grip with 2 foot switches for the bot's gadgets -- that would be the way to go!
Just because the device the human is using to control the "robot" is a computer and the "robot" has a computer onboard does not make it a robot.
With the cheap processing power available today and the current state of AI there is no excuse, bar incompetence, for this competition to not consist of truly autonomous robots. Until then, Battlebots will continue to be a show pandering to the lowest common denominator, relying on sex and loud music to attract an audience.
I've submit this cool-looking, Debian GNU/Linux powered battlebot TuxBot but got rejected by /. Gods.
Now this rugged oranges box got spotlighted?
What's wrong with you guys?
and that problem is that when your "Doom-bot" runs up against your competitor's $4000 titanium shell that laughs at whatever sad-assed weaponary you have. There is a reason that the 'bots (at least the ones that win at least once) on Battlebots cost a shit-load of money- quality parts. Yeah, I could stick a stuffed monkey in a Barbie car with spikes on the front, but I think that the kill saws and the hammer will make short shrift of said device. This ain't "lawnmower vs. weed-whacker" kids!
Vote monkeys into Congress. They are cheaper and more trustworthy.
I'm wondering how much time (not to mention cpu cycles) these guys could have saved if these guys had hacked a much simpler operating system ? Perhaps using some old-school PD source such as Tom Poindexter's Crobots or perhaps or one of it's various mutant prodginy such as the linux based C-Robots ?
healyourchurchwebsite.com - WWJB?