Slashdot Mirror
AIBO Robot Dog Soccer Competition
BrianWCarver writes "The Washington Post has an article about teams of college students who program Sony AIBO Robotic Dogs to play soccer against each other in teams of four. While Beckham's job is not yet jeopardized, the cool thing from an AI perspective is that 'once the humans flip the switch, the robots are on their own.' They compete in RoboCup whose stated goal is to 'by the year 2050,
develop a team of fully autonomous humanoid robots
that can win against the human world soccer champion team.' RoboCup also has competitions with wheeled soccer bots (of varying designs) and have a humanoid league in which the Honda ASIMO appeared. The students in the above article are preparing for the four-legged international championship coming up in July of 2003 in Padua, Italy."
In this earlier story we are told that AI is going no where due to the fact that graduate students are spending 3 years fixing broken robots instead of programming them.
I think if they give up those big robots, buy some Aibos in bulk, they can skip the repair stage, or at least shorten it quite a bit. Of course, if they get them at Circuit City, they can get the extended guarantee and not have to worry about repairs at all!!!
---
Never call a man a fool; borrow from him.
R2D2 can now be a proud soccer mom
Im dreaming ofa big bndwdth, That can resist the
When a humanoid soccer-playing machine can be developed such that a simple slide tackle (a legal maneuver) can snap limbs, the world cup is a sure bet. This will take far less than 47 years. In fact, drop the "humanoid" requirement and we could be there next year.
Sometimes there's an elegant beauty in simple brute force.
This is a very ambitious goal for the RoboCup team ... it will require great strides in so many areas: things like image recognition, mechanical and electrical engineering, and a severe amount of artificial intelligence breakthroughs. Soccer is not just a game that can be "solved" like checkers, tic-tac-toe, and awari are, and chess will be. It requires a much higher level of artificial intelligence (decision making, goal-based planning, etc.)
:)
It would be very neat to see something like this happen, and I know 47 years is a long time, but it's still an incredibly ambitious goal
the blood has stopped pumping, and he's left to decay
the me that you know is now made up of wires
7. This will get the attention of some geeks; maybe enough to draw attention away from "Star Trek" and cause this dead franchise to finally go away.
6. The moment when the goaltender says "Danger, Will Robinson!" is priceless.
5. Lots of fun game play alterations when robot bones are tossed onto the field.
4. Only a matter of time before they allow Taz-Bot, Die-Sector, Vlad, and the rest of our Battlebots friends onto the field. Then the fun really starts.
3. Gets rid of the problem of urination on the field that is plagues any event with real soccer players or real dogs.
2. Americans love dogs. Americans hate soccer. So, this is likely to move soccer from #74 to #63 on the ESPN-2 ratings.
1. Two words: Mecha-Aibo
Don't blame Durga. I voted for Centauri.
How else do you hold an "AI robot" contest? Humans mucking about with the things is called cheating. If they get disabled or whatnot, of COURSE they should be left alone.
I hate robot/AI contests which are dumbed down- watch a robot 'soccer' match, and often you'll see volunteers putting robots back on the right course when they've boxed themselves into a corner and such...like the programmers/designers shouldn't have to be 'troubled' by such things as getting trapped by two walls, or all the contestants have such miserably designed/programmed robots that they fail left and right.. Everyone wants to work on the "chase the ball" routine, but nobody wants to work on the un-sexy, nuts-and-bolts, "keep from smacking the wall and staying there" routine.
In the real world, there are no magical hands that pick you up and flip you around and set you going with a pat on the CPU...and what every robotics person calls "simplifying the problem", I call "cheating". This constant cheating has led to a field which is incapable, still, of dealing with the simplest problems but can solve these wonderful complex ones. The result is a lot of electromechanical garbage that's simply unuseable in anything even remotely resembling the real world.
Please help metamoderate.
10. Drooldroids
9. Dyna-Mutts
8. Cyberpups
7. WinterMalaMutes
6. FIDO-Net
5. 3L33T T3RR313R's
4. The B-9 K-9's
3. See-Threepoodle's
2. The Borgzoi
1. BITches
Don't blame Durga. I voted for Centauri.
This competition is similar, although in my opinion has more engineering merit. I spent a year developing a team for the 2002 Korea competition but unfortunately we were beset with a poor budget (NZ$5000 - things got tight) and outrageous shipping delays (6 months for motors) and never completed the team.
It was interesting because not only did you have to develop the AI to allow the robots to 'play the game', but you also had to develop a computer colour vision system to 'read' the state of play, as well as a suitable control system for the robots themselves. The use of H bridges and avoiding burning out the motors or circuitry when suddenly reversing direction brought in some interesting research from the university's mechanical engineering department. Wireless comms also came into it, with a one-way FM link. A great project bringing many different areas into one 'arena'.
FIRA has several different classes of competition (we were working towards Mirosot) including a 'simulation only' class.