Slashdot Mirror
BattleBots Delayed, Will Go Brains Over Babes
An anonymous reader writes "We got all excited earlier this week about robo-combat returning to TV with ESPN, but now PopMech super geek Erik Sofge talks to the folks at BattleBots and finds out that because of so many early entries, the competition will be delayed until at least November. The reason? Gone are the babes and predictable wedge fights, in are eager engineering students, a crazy ramp arena and lots of new rules. Worth the wait, or do we miss the Comedy Central version?"
First off, what did Comedy Central have to do with fighting robots? ESPN I sort of get (they call Poker a "sport", so I guess battlebots can be one, too), but why Comedy Central.
More to the point, I miss Battle Bots being televised, but I don't miss Comedy Central. The announcers were annoying, Carmen Electra was pointless, and they spent too much airtime dithering about nothing. The worst of it was that the camera angles and microphone pickup made the bots look like toys. Many of these machines took a team of guys to lift them out of the travel van and get them into the arena. Every year, they thickened the lexan around the arena, and every year, something managed to pierce it. These are nasty machines, but they never looked more powerful than an unmodded nerf gun.
Not a typewriter
While I agree that it does seem kind of lame that a show titled "BattleBots" featured no actual robots, I highly doubt that a show featuring autonomous bots would actually be interesting to watch.
To borrow a quote from bash.org, all you'd end up watching are robots that "collect data about the surrounding environment, then discard it and drive into walls."
But even assuming that the AI could match the human-controlled bots, it wouldn't make it any more interesting. Instead of watching a bunch of human-controlled wedges attempt to flip each other over, you'd be watching a bunch of computer-controlled wedges attempting to flip each other over. Intellectually more interesting, I suppose, but really no more fun.
The real problem with BattleBots is that it's just boring. I don't want to watch a bunch of robots attempt to score points on each other and then watch the judges declare one the winner, I want to watch robots destroy each other. In BattleBots, the destruction was generally limited to "something's bent and it no longer moves."
I doubt BattleBots will ever really be all that interesting. Ultimately it's going to remain a bunch of bots hitting each other until one craps out or time runs out.
Hopefully I'm wrong and ESPN will surprise me, but BattleBots strikes me as one of those concepts that sounds like fun to watch but proves not to be, regardless of what actually controls the bots.
You are in a maze of twisty little relative jumps, all alike.
I absolutely loved Robot Wars when it was shown here in the UK, one of the tv hilights of the week for me, seeing machines bash the shit out of each other is much more entertaining than any boxing match or wrestling match.
I'd heard about Battle Bots in the midst of watching Robot Wars and eventually one of the tv stations here aired a few shows, what a complete let down it was! there was some serious hostility of the opponents being portrayed on screen, wether it was real or staged it just felt nasty compared to the UK's Robot Wars where the teams would help each other out because they were mainly there for the whole entertainment & experience, not to gloat over their wins.
I would hope there's more friendly rivalry in the upcoming Battle Bots instead of nasty "I'm better than you" type taunts & attitude, but being an American show I don't hold out much hope, except perhaps that someone will torrent the shows because I really miss seeing robots beat the crap out of each other, half because of the fights and strategy involved and half because I love seeing the ingenuity and designs of the robots.
To do something right, you often have to roll up your sleeves and get busy.
some of my favorite matches had only-a-wedge-on-wheels bots! it was funny watching them take one or two hits from a pneumatic pick hammer and stop working, or even better, get flipped over and impaled again.
Why is it that someone always feels the need to say this on EVERY SINGLE BattleBots related thread?
Your design to a real part online: Big Blue Saw
Trouble was, the robots that rivalled Hypnodisc for the title were always well built and well armoured. So Hypnodisc would utterly destroy the no-hopers, and then when it met the likes of Cassius or Chaos or Razer or Panic Attack it found itself rather worse off. A flywheel weapon relies upon the one-hit kill; it's expensive in energy, so we often saw Hypnodisc running low on power after a few jolts, and left vulnerable. Worse yet, hit something that doesn't yield and you hurt yourself maybe more than the target...
Real Daleks don't climb stairs - they level the building.
The Comedy Central version was just pathetic. At first it started off talking about the robots, but slowly turned into a joke. I blame both the contestants and Comedy Central. I blame Comedy Central for taking the show and taking away the nerdy aspects and instead focusing on marketing the show to more people and they did a pretty bad job at that with my next reason.
I blame the contestants for not letting their robots get smashed to bits. Most battles ended with the winner giving pity to the other contestant. Only a few battles featured the other robot getting destroyed. I wanted to see the robots fight to the death. Its like the money shot of the show. They first discuss the robot and how it works, all the aspects behind building it, etc. Then they go at it. I know alot of hardwork go into building these but seriously, when you put your machine in a arena full of dangerous obstacles and a competing robot, do you expect it to come out peacefully.
The majority of battles ended with a robot malfunctioning on the first blow and then if it didn't work after 30-60 seconds they ended the round.
Given where student robotics already is, and given the software availability, what more can you possibly need for a perfectly viable autonomous robot league for Battlebots?
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
- Bonus mass allowance for autonomous robots. Remote control is neat and all that, but real robots have senors and logic (and flamethrowers)
- Bonus mass allowance for legged robots. Additionally, there should be terrain features that favor legged bots over wheeled bots. Steps and potholes come to mind. Potholes with flamethrowers.
- Large bonus mass allowance for bipedal legged robots. Hey, if you can make a bot walk and smash things then you're doing alright. If it walks and carries a flamethrower it is just about ideal.
- A venue where flame and projectile weapons are safe (and encouraged). The ultimate bots from the previous incarnation of the series tended towards those that stored up huge amounts of angular momentum. Without the ability to counter this huge offensive potential with countermeasures such as nets or chains or flails these bots were almost unstoppable. I'm not sure how a flamethrower would help here, but flamethrowers make for great TV.
Finally, using the aforementioned flamethrowers, get rid of the WWE style announcers.Peter
Downsize DC Today!
you are assuming that the AI is a true AI,
No, I'm not. I'm talking about "weak" AI, the practical side of machine learning and decision making algorithms. Look at our best robots -- Asimo, the RoboCup, even industrial machinery controllers -- and you'll find systems that operate between passably and impressively in their particular controlled environment, but are fairly useless outside it. Even RoboCup -- soccer playing robots -- only works because all the robots are essentially the same, and the ball and nets are constant, easily identified objects, and the objectives of the match are well defined.
That's what I'm saying, it would work, but it'd have to be a completely controlled environment without the physical creativity of BattleBots. That would still be a cool thing to see, but it would push the competition almost entirely into the realm of programming. Which isn't as much of a spectator sport, even if geeks like us still think it's awesome.
using detailed radar or video analysis you can search for fast moving parts since most weapons will be fast movine, then automatically evade what appears to be the range of motion of said moving part. sure it's more complex than a roomba but it's not impossible.
And until it moved, it would have no idea it was a weapon, and when it stopped moving, it'd lose the weapon, unless it was programmed with algorithms to identify that particular kind of weapon.
Of course it's not impossible to do all this at some point, and it will be sweet. Like I said, I'm not assuming "strong" AI which we may never create because we don't even know what it would look like, it's more "weak" AI that is really a matter of engineering and algorithms. I just don't see that we are there yet. For now, I think the radio-controlled but creatively designed "robot" combat is cooler, because not only do you see innovative designs and ideas for ways to build robots, there's room for skilled and adaptive piloting to make a difference.
The enemies of Democracy are
Engineering students will find the problem of autonomy more interesting, but when it comes down to it, human control is going to give a team the greatest chance of winning. Furthermore, if autonomy is not required, the time spent developing it would be better spent on weapons and defense systems.
The only way autonomy would happen is if it is required. I'd love to see it as a requirement, but the fact is that controlled machines are going to be more interesting to watch, even if they represent less of a technical achievement.
Perhaps they could have two subclasses, or have the robots fight one round autonomous and one round remote controlled. I also think each team should be required to release their AI code under the GPL at the start of each competition. That way, after every competition the AI will keep getting better, and new teams won't have to start from scratch.
New rules :
* the first day, every robot fights a simple robot or remote-controlled one, created by the organizers. Those who can fight it are admitted. These challenges are public so that each team can see the other teams' robot in action.
* 24 before the fight, teams know which robot they will have to fight
* They have one hour between each battle to change the program of their robot or to input new parameters
That's approximately what is being done in the European Robotic Cup Eurobot. The intent is not the same as the battle bot : here the robots are in the same arena but must not fight each other, they must accomplish a task more efficiently than the other, possibly by stealing some of the opponent's items or by moving in a way to annoy it. There is a children version of this challenge that involves remote controlled bots, but that's considered as an uninteresting challenge by roboticians.
Some robots are designed in order to detect the other robot and adapt its strategy accordingly. Most challenges are designed so that a "dumb approach" (with no or few sensors) is still possible but very inefficient.
The Wise adapts himself to the world. The Fool adapts the world to himself. Therefore, all progress depends on the Fool.