OK, I admit it – I’m fond of the TV show BattleBots (now on for a limited [I think] run).
But every time I watch I’m irritated by the name, since the fighting machines aren’t IMO robots at all, but fancy weaponized remote control cars.
So how come there aren’t any real BattleBots – machines that wreak mayhem on each other without a human operating them? Given the advances in technology this would seem to be possible.
Possibly because having a TV show in which one battlebot tries hopelessly to figure out how to get out of a corner, while its opponent wanders aimlessly around the arena looking for it, doesn’t make the best television. AI robot competitions are fun and interesting but probably not for the mainstream audience, and generally involve teams of college students competing to perform very specific tasks.
While HP or Google given a few years could probably come up with a battle bot that would kick ass, I think don’t think a father and son team is likely to be able to put something together that can handle all of the complexities of a full battle situation.
That certainly makes sense. Which leads to another question – why don’t Google and HP sponsor true battlebots? Almost all of the current contenders have corporate sponsors, and it would seem to me that the companies could learn a lot that would help in money-making ventures (such as self-driving cars).
In fact, it seems to me that self-driving cars would be more complex than battlebots. The bot has to deal with only one other vehicle, while a car has to deal with other vehicles operated by drunks or idiots who are too busy texting to pay attention or exhausted folks who fall asleep, etc., etc.
For that matter, why wouldn’t the DOD sponsor a couple of bots in the hope of coming up with some good ideas for future weapons?
Back in the early days of robot combat events, before it was televised and big-budget, there was an experimental autonomous class. These matches mostly resulted in several minutes of robots trying and failing to find each other in the arena, until one of them random-walked into a wall or arena hazard and got stuck.
Real autonomous battle-bots exist. I’ve built one, and officiated over a tournament between them (mine lost, but then, I was hoping that everyone would figure out how to beat it anyway). We used Lego, which frankly seems to be the medium of choice for most amateur robotics nowadays.
Since Google is trying to convince people that their cars won’t kill them. This is probably a bad idea from a marketing standpoint 
I have a family member who participated in autonomous robot battles and that was, gosh, probably 20 years ago. He was a robotics engineer at the time though so I don’t know if it was an open event or just something competing labs did for fun or what.
“Hey General, I think your multi-billion dollar weapons research programs might be missing out on some good ideas. I think some backyard mechanics with no budget fooling around with toys for television shows might have better ideas than the guys you’re using now.”
[QUOTE=Anny Middon]
For that matter, why wouldn’t the DOD sponsor a couple of bots in the hope of coming up with some good ideas for future weapons?
[/QUOTE]
You could ask iRobot. For the home, they make the Roomba, and for the military, they make robots that can do surveillance and bomb clearing, as well as carry weapons, in a range of sizes from a five pound “tossable” up to one that can reach up to eleven feet.
Surely, any robotic vehicle would have to carry all kinds of sensors, just like a Google car. Knock one out and the robot is just a dumb chunk of machinery.
We had a very similar series over here. When it was just family teams building them in the garage, it was fun. When the robots started to get sponsorship and machine shop made parts, it lost its appeal. They tried putting a cost limit on them, but it was too complicated and the audience switched off.
Possible, but damned expensive and time-consuming. Computer vision is still very much a research area. So you end up with a mix of fairly crude sensors that can detect barriers and obstacles and maybe some more sophisticated sensors for moving objects. But analyzing the moving objects takes a lot of computation time which you don’t have in competition, because someone is trying to slice your chassis off.
And, if you succeed in creating an autonomous battlebot, it would look, to onlookers, exactly the same as it does now.
A hybrid approach might work, though, with an autonomous weapon system that under certain circumstances can operate faster and more precisely than human reflexes – something like autofocus in modern SLRs.
You mean like this?
Today at camp, we decided “To heck with it. All the campers want to build battle-bots; OK, we’ll build battle-bots”. They’re a lot more crude than ones with a week of design time behind them, but they are autonomous.
Believe it or not, but sometimes amateurs can explore the design space more throughly and cheaply than overpaid PhDs. Amateurs can get access to equipment in the same rough class - good quality sensors, powerful computers with lots of memory - as the professionals. Sure, the professionals have better stuff, but it may not be so much better that it matters.
I seem to recall watching this on PB **several **years ago, and enjoying it greatly.
duplicate post, nevermind
We’re getting close to that actually. Between the Robot Combat League ( Robot Combat League - Wikipedia – basically exo suits) and some smaller fighting robots (https://www.youtube.com/watch?v=j4zTOfix40Y) I’m guessing a pre-programmed robot fighting without a human operator may be within 5 years or less.
What’s the over-under from then to the Great Murderbot Rebellion ?
That’s probably why its taken so long as it is. According to some of the CMU lab rats I go drinking with now and then, between drone development and other advances, basically all the pieces for some sort of true fighting robot are in place somewhere; they just haven’t been brought together at this point. Humans being human, we want to be able to point to some person and say “you screwed up” when the wrong target gets hit or we have “collateral damage” as innocent deaths are called these days. So what is being done is almost all in terms or remote operation rather than autonomous functioning. But ----- with what is being developed for driverless cars ----- apply it a little sideways and you get something that almost sounds frightening when developed by the government. Let some sharp engineering geek do it in his garage and we’ll probably be OK with it.