University of Pennsylvania
Legged RoboCup Soccer Team

The UPennalizers, the University of Pennsylvania's Robocup Team participates in the annual Robocup legged soccer challenge. We have competed in RoboCup '99, held in Stockholm, Sweden, RoboCup '00 held in Melbourne Australia, and RoboCup '01 (see the legged competition) in Seattle, Washington, USA and RoboCup '02. We participate in the RoboCup Soccer Legged League, which uses Sony's legged robot, the Aibo.

For the 2001 competition, Penn's team came in third out of sixteen teams in the main competition, and second out of sixteen in the RoboCup Challenges (technical challenges).

Next year's competition will be in Padova, Italy.

We are currently in preparation for next year's competition.

Some helpful documents for compiling sample programs and using the Matlab interface

• Compiling Sample Programs using the OPEN-R SDK
• Using the Matlab interface

Some useful links related to this project:

• General RoboCup Website - General information about the Robocup competition
• 2003 RoboCup Legged League Website - This site has specific information about rules and regulations for the legged league competitions. Also, you can access technical reports from other teams through this site.
• 2002 RoboCup Movie Clips - Movie clips from Robocup 2002.
• Sony's OPEN-R Website - You can download the OPEN-R software development kit, as well as documentation for the SDK here.
• 2003 RoboCup New Rules and Proposed Guidelines - See the new rules and tentative technical challenges for next year's competition.

We've written a couple papers on various aspects of RoboCup:

• Our software architecture (including MARS project)
• A paper on Omnidirectional Walking
• A for the RoboCup book, describing our team's involvement in the 2001 competition.

Here are also some movies of the dogs in action (they don't seem to play on UNIX based mpeg players, so good luck...).

• A side kick and long goal (3.5MB)
• Our stellar goalie save in the RoboCup Challenge (8 MB)
• A goalie save and dribble to goal (8 MB)
• Demonstrating a potential field-based obstacle avoidance behavior (2.7MB)
• Demo of collision detection and deadlock resolution (detecting a "stuck" condition) (2MB)

Here's a good shot of our team from the Seattle competition (courtesy of Glenn Zucman):

And here are a few more photos (courtesy of Kentaro Fukada and Sony):

Also, here are a few old movies for comparison [movie1], [movie2], [movie3],

Last updated October 16, 2002.