| |
|
UvA Trilearn 2001 - Soccer Simulation Team
Objectives
'UvA Trilearn 2001' is a team that participated for the first time
in RoboCup 2001. The team had been built from scratch and did not contain any
code copied from other RoboCup teams. Therefore much of the initial effort had
gone into getting the lower level aspects to work. The intention was to try
to improve upon some of the low-level methods used by other top teams from the
past. This has among other things led to an advanced synchronisation scheme and
very accurate methods for position and velocity estimation. Much attention has
also been paid to the software engineering aspects of the project for
comprehensibility and to facilitate future use. Long term scientific goals will
focus more on higher level aspects such as multi-agent strategic reasoning,
adversarial planning and multi-agent modeling using a coach agent.
Research Group Members
dhr. J.R. Kok (jellekok@science.uva.nl) -
Research and Implementation
dhr. R. de Boer (remdboer@science.uva.nl) -
Research and Implementation
Prof. dr. ir. F.C.A. Groen (groen@science.uva.nl) - Team Coordinator
dr. N. Vlassis
- Research
Coordinator
Funding
University of Amsterdam (UvA), Faculty WINS, Department of Computer Science
Intelligent Autonomous Systems Group
Kruislaan 403, 1098 SJ Amsterdam, The Netherlands
Key Publications
- An article about our graduation project appearing in the WEIS faculty paper
'Dirk' in april 2001 (only in text format and in dutch):
Robot Voetbal Simulatie
- A general team description of the 'UvA Trilearn 2001' soccer simulation team
which describes the state of affairs in april 2001:
TeamDesc.apr2001.ps.gz
- A research abstract describing the research focus of the 'UvA Trilearn 2001'
soccer simulation team (this paper was part of the qualification material for
RoboCup 2001): ResAbs.apr2001.ps.gz.
- An article in the BNVKI newsletter
describing some global aspects of our team and the results of our team in
Seattle: BNVKIarticle.280801.ps.gz.
- Team Description Paper for RoboCup 2001 book: TeamDesc.sep2001.ps.gz.
- Towards an optimal scoring policy for simulated soccer agents (technical
report): IAS-VA-01-06.ps.gz.
- The Incremental Development of a Synthetic Multi-Agent System:
The UvA Trilearn 2001 Robotic Soccer Simulation Team. Master's thesis:
thesis.ps.gz,
thesis.pdf.
Binary
The binary of UvA_Trilearn 2001 as used during the Robocup in Seattle is
available here.
Source
Parts of the source of UvA Trilearn 2001 are available
here. The released source
contains our lower levels (synchronization, world model, basic agent skills)
together with a simple high-level strategy, similar to the one released by
FC Portugal after RoboCup-2000 to make a working team. This code is released
under the modified
BSD-license. This basically means that
the software may be used, modified and copied as long as the copyright notice,
the list of conditions and the disclaimer are reproduced in the documentation
and/or other materials provided with the distribution.
The associated html documentation can be generated from these sources or
be downloaded here.
Competitions
- German Open
We participated in the
German Open 2001 and became 5th (of 12)
in this competition.
Results:
| Round | Opponent | Score |
| Group | Aras (Iran) | 2-2 |
| | FC Portugal (Portugal) | 2-5 |
| | Lucky Luebeck (Germany) | 4-0 |
| | Sharif Arvand (Iran) | 1-2 |
| | Osna Ballbyters (Iran) | 16-0 |
| Elimination (winners) | Dr. Web (Russia) | 0-1 |
| Elimination (losers) | RoboLog (Germany) | 4-0 |
| | Dr. Web (Russia) | 0-1 |
| | | 29-11 |
All results can be found at
http://www.inb.mu-luebeck.de/robocup/GOResults.html.
- Robocup 2001 Seattle
We participated in the World Championships
RoboCup-2001
in Seattle and became 4th
(of the 42 qualified teams) in this competition.
Results:
| Round | Opponent | Score |
| Group I | TsinghuAeolus (China) | 0-7 |
| | Living Systems (Germany) | 6-0 |
| | Harmony (Japan) | 3-0 |
| | rUNSWift (Australia) | 2-0 |
| Group II | TUT-groove (Japan) | 6-0 |
| | A-Team (Japan) | 4-0 |
| | FC Portugal 2000 (Portugal) | 6-0 |
| | Cyberoos (Australia) | 8-0 |
| | Rolling Brains (Germany) | 5-0 |
| Elimination (winners) | Wright Eagle (China) | 2-1 |
| | FC Portugal 2001 (Portugal) | 1-4 |
| Elimination (losers) | FC Portugal 2000 (Portugal) | 6-1 |
| | Brainstormers (Germany) | 0-1 |
| | | 49-14 |
All results can be found at
http://www.cs.cmu.edu/%7Erobocup2001/AllSchedules.html or
here.
More Information
More information about the RoboCup Initiative itself can be found at
http://www.robocup.org
The official site for the German Open is located here
The official site for Robocup 2001 is located here
Qualification Material
The qualification material for RoboCup 2001 can be found below. We understand
that all binary products participating at RoboCup 2001 soccer simulations,
whether for player clients, coach, and/or other tools used in the competition,
are REQUIRED to be released for research usage by August 31st, 2001. We also
understand that source code release is highly encouraged.
- A game log of 'UvA Trilearn 2001' playing against 'FC Portugal' using server
7.xx for a single full length game:
qual_logfile_uva2001.log.gz(3.5MB).
Note that this logfile (version 2) was
created using soccerserver 7.04. It has been generated on two AMD Athlon 700
MHZ, 512 MB machines running Debian GNU/Linux with kernel 2.2.17 which are
connected by a 100 Mbit switched ethernet network. More information about the
hardware used can be found here. The
soccerserver/monitor and our own team 'UvA Trilearn 2001' have been run on one
machine and 'FCPortugal' on the other. If necessary the server logfile containing
all the commands is also available:
qual_server_uva2001.log.gz(2.7MB).
- A 2-page research abstract describing the research focus of the team and
an overview of results in RoboCup and any additional domains: ResAbs.ps.gz.
A remark is in order here. At first we thought that this meant that we had to
write a team description paper containing a general description of our team. It
was only several days before the qualification deadline however when it was
clarified to us what was actually required. We then transformed our team
description paper into a paper that better conformed to these requirements and
the result of this can be found above. We strongly encourage however to also
read our original team description paper (which can be found here) since it contains more complete
information about our current implementation and future plans.
|
