[robocup-legged] Rule Changes and Poll

[Thomas Röfer]

Hi,

> I am afraid that under the current limitation of the Aibo's vision
> the variant "3) no beacons" (although very appealing) incorporates
> the highest risk of not to achieve the goal of major innovations and
> high quality games in 2007.

I think there is a certain misperception about what makes our league
interesting in the eyes of the RoboCup trustees. It is basically the
research that is performed and how the work done contributes to the overall
goal of RoboCup. There is surely a certain interest in high quality games,
but it is most certainly not the intention of the RoboCup trustees to stop
the continuous change toward the 100x70 qm outdoor field in a soccer stadium
the 2050 humanoid robots will have to deal with. With removing the beacons
the quality of game play may be reduced in 2007, but I would not expect it
to drop dramatically.

We need new challenges. In fact half of the talks given at the RoboCup
Symposium 2006 were coming from our league (congratulations!!!), but I had
the impression that most of the things presented did not make its way to the
actual games. For the AI magazine article on RoboCup 2006 I had to answer
the question "What are the improvements from 2005 to 2006?" and I had no
good answer to this question, for instance I found no answers in the TDPs of
the two finalists.

Removing the beacons may ruin self-localization. So either we have to invent
methods to still be able to metrically self-localize without them (I think
that is still possible), our we have to rely on a more qualitative way of
self-localization (as human soccer player most certainly do). So this poses
challenges to most parts of the control software: image-processing (perceive
more/other objects), world-modeling, behavior control (play with more
uncertainty, without metric position), attention control, and multi-robot
collaboration (can teammates help the attacker in localization?).

After 2007, we may change the robot platform to something new. Since it is
quite sure that the central sensor will still be a directed camera, the
research on localization without beacons is not in vain.

Best regards

Thomas Röfer

_______________________________________________________________________
Dr. Thomas Röfer                   Office Address:
DFKI-Lab Bremen                    Universität Bremen
Safe and Secure Cognitive Systems  Cartesium 0.055
Robert-Hooke-Str. 5                Bibliothekstr. 1
28359 Bremen, Germany              28359 Bremen, Germany
http://www.dfki.de                 www.informatik.uni-bremen.de/~roefer

Phone: +49 (421) 218-64200
Fax:   +49 (421) 218-9864200
eMail: Thomas.Roefer at dfki.de