Slashdot Mirror

Ok, I'm not there but as far as I know android team v. android team is not being done this year, contrary to what the article is insinuating.

According to the Official 2002 Robocup Humanoid League Draft Rule there are three catagories of Current and Future events with several sub-sections. Here is the run down.

* Standing Still on One Leg
* Humanoid walk - out from one end, around a pylon and back,
* Shoot - where the bot is able to shoot on the goal and get it in.
* Penalty Shootout
* 1v1, 2v2, 3v3 Soccer
* Freestyle - Five minutes of judged performance art.

According to the organizers they are just hoping to get some teams to try the first few! And as you can see the competitive playing of soccer is waaaaaaaaaaaaaaayy down the list and probably aren't even being attempted this 'Cup. Sure, those nice posed shots of those Sony bots look nice, but those aren't competition pictures. Sure, the information on the Official Robocup site is sparse, but don't you think that they would have some big announcement if it did?

This article from MSNBC is confusing one league (the humanoids) with another (the non-humanoid) in an attempt to create hype. Personally I think Robocup deserves it, but not by misconstruing what's going on.

The best thing about the Robocup site is that you can actually watch the replays of the simulation games with Flash. Pretty sweet.

Well, From June 19 through the 25th, we can have a daily article or three on World Cup as long as we tie it in to RoboCup stories. What I'd like to see on the front page are the daily scores and stats from THAT. (I'll get my soccer stats live and from sources other than /. thankyouverymuch.)

What else would you expect Germans to teach little Aibo dogs? - Soccer of course!

The World Championship this year is to be in Fukuoka Japan.

Have fun!

How does one compete in this tournament? It looks like all the teams have identical equipment: do the companies (like Sony) sponsor them?

There are several different leagues that one can enter. There is a simulator league, small-size league, middle-size league, Sony legged league, a humanoid league (new for this year), and a couple rescue leagues.

The simulator league has no hardware. The software is downloadable from sourceforge. This is the easiest league to get started in since there isn't a lot of money involved and you can even enter remotely. Keep in mind that a lot of researchers are working hard on this problem, so most of the teams are really good. The Sony legged league is run by Sony. Sony provides a lot of support. You have to make a proposal to Sony to get into this league. A committee of researchers and Sony personel decide which teams to allow in. Everyone uses the Sony hardware in this league. The small-size and middle-size leagues have hardware built by the teams. It usually costs at least $10,000 for all the hardware needed for a small-size team. Middle-size teams run from $3,000-$40,000 per robot (4 on 4 competition). See www.robocup.org for more details. Registration is already over for this year, but there will be another competition next year (in Italy, I believe).

I don't know what Mr. Blach is exactly doing but the Robocup is already there for years.

George Jetson had items like a videophone, flying car, and a robot.

You're basically swapping sub branches of your program around to see what works - tranversing the space of all possible programs - it takes *a lot* of random attempts

This is quite a good summary of GP, and it also describes why IMHO GP should be the last thing to try. i.e. use it when everything else has failed...

I think reinforcement learning (RL) methods are much more promising in general, and should be tried before GP whenever they are applicable.

In GP you set up a fitness function that evaluates the performance of the entire program, while in RL the reward function rewards individual actions i.e. small parts of the program. The big problem in both cases is the design of the fitness/reward function. In GP you want it to increase fitness as your agents grow better and better (yes this is difficult, I tried it once in the RoboCup simulation league). In RL you can give rewards very easily by basically saying good/bad to the program after it has performed an action. The problem in RL is that the program has to find out which of the recent actions caused the reward. There is a large number of algorithms for distributing rewards among recent actions. However, they all do better than GP, since GP is equivalent to the extreme case of just one reward after the program has finished running.

href="http://www.r.cs.kobe-u.ac.jp/robocup-rescue

you can build your own coordinated team of virtual robots to work in a simulated disaster zone + there are some search and rescue competitions using real robots. There's a download page for the simulators and disaster toolkits.

http://www.robocup.org/

For those of you interrested in trying this at home (yes you can :-) the software is available for download. You can run the "server" on both Unix and Windows. Its just to write a few players and get rockin!
For those who just wants to see how it all works, there are also old teams to download. Go fetch!

• mayhem is more fun than constructivity. Contructivity can be more interesting for those who have a special interest int the technology behind
• mayhem is easy, you don't need no brain to run around trashing opponents

The Robot World Cup Initiative (RoboCup) is an international research and education initiative. It is an attempt to foster AI and intelligent robotics research by providing a standard problem where wide range of technologies can be integrated and examined, as well as being used for intergrated project-oriented education.

For this purpose, RoboCup chose to use soccer game as a primary domain, and organizes RoboCup: The Robot World Cup Soccer Games and Conferences. In order for a robot team to actually perform a soccer game, various technologies must be incorporated including: design principles of autonomous agents, multi-agent collaboration, strategy acquisition, real-time reasoning, robotics, and sensor-fusion. RoboCup is a task for a team of multiple fast-moving robots under a dynamic environment. RoboCup also offers a software platform for research on the software aspects of RoboCup.

This is not a TV-show, but maybe in twenty years the technology will reach a point where the robotic players are skillfull enough to produce games that are exciting for the masses?

Not so! Check this out. Most robot competition is non-violent. Though it is fun to see parts fly.

No, this is not a joke ;) Sony is giving away, at cost sony legged robot research prototypes for, of all things, a international robot soccer competition to spur research into artificial intelligence and robotics.

Have some under-utilized grad students in a computer science department? Get them to work on preparing research software, [and get some local sponsership in the neighborhood of (I'm guessing roughly) [15,000 - 25,000] dollars, and you may get to have some real fun to justify all those boring years spent in stuffy classrooms studying how to convert certain classes of NDA's into DFA's !!!

read about it here


---
man sig

the firefighting robot competition at [snip]

RoboCup-Rescue Simulation Project is a new practical domain of RoboCup. A generic urban disaster simulation environment is constructed on network computers. Heterogeneous intelligent agents such as fire fighters, commanders, victims, volunteers, etc. conduct search and rescue activities in this virtual disaster world. Various people worldwide participate in this simulator as an entertainment, a training, or an education via the internet. This project contributes not only to enlightenment of citizens but also to accrate damage prediction, decision support in real disaster, systematic real time monitoring of actual damages, and emergence of better disaster prevention strategies. A diverse spectrum of possibilities of this technology will contribute to the creation of the safer social system in the future.

though I wish I could find official rules somewhere

[snip] that these are controlled by an off-field computer but at least their behavior doesn't seem to be modified by humans at runtime.

So, goodbye Marvin Minsky! So long, later John McCarthy! We'll see you in the Open Source AI!

This is a very good idea that should be adopted more widely I think. (Things are moving in this direction already though. RoboCup has helped a lot to get people started.) In today's AI, there are so many people working on inventing new programming structures or ways of arranging "knowledge", without ever bothering to think about how they should obtain this knowledge. If they had the programming structure given, they could concentrate on trying things out on real problems. The AI community suffers from way too much theory way too little practice IMHO.

I'm currently working on such a thing to learn soccer strategies for RoboCup. The general idea is to couple feedback type reinforcement learning with a large, distributed tournament-style evolution system.

"Take the example of something more fast-paced than Chess like Soccer."

Check out http://www.robocup.org , a robot soccer competition where autonomous robots play team soccer.

For 'team' robotic efforts, one resource to check out is Robocup, a world-wide robotic soccer competition.

I don't have any specific references here, but I believe there have been some published papers (as a result of this competition) regarding team coordination and strategies. A good starting search would be the various AI journals.

Good luck!

Any other ideas for good AI benchmarks?
The benchmarks have to be there to encourage funding and some research, so some test needs to be decided as a standard.

How about RoboCup?

Computer vision is a decent test, but it has to be under such tight constraints. Other senses aren't worth the time, so we are left we only a few options.

This remark leaves me clueless. What's wrong with the vision problem? If you have a computer system that can see, wouldn't that be useful?
What are those constraints? If you are referring to the great need of processing power, I'd say thats more of a challenge (if the algorithms require too much processing power, then maybe they need rethinking)!

If U prefer smart bots, you can watch soccer games between teams of bots with real AI.
There's a world cup each year.
Check out robocup.org

In a related note, there's currently the robocup going on over in Sweeden, and
there's a Sony Legged Robot League which is these AIBO's running around playing
soccer.

It's quite funny when they fall over and lay around flapping their legs till
they manage to get up.

There's a RealVideo stream here
More info about the competition and game schedules are over at www.robocup.org

What's really cool is that the Sony AIBOs were actually given their own league in this competition. They really can do more than just get up when they fall down.

Carnegie Mellon University has been winning most of these competitions in the past few years - check out the team's project leader's web page here.

Yup, they've got a whole league running as a subsidiary of IJCAI. For on the robocup more see the homesite.


|--Arzin Tynon---

These robots can be viewed playing soccer in Stockholm (Sweden) this summer. A number of universities around the world will compete against each other. This event is called RoboCup.

... just not in the commercial market. This robot has it's own league in the RoboCup Soccer Tournament. The winners last year were a team from Carnegie Mellon University (go CMU!! - I go there). It's really cool. Here's the site of CMU's team leader:

Manuela Veloso's Page