Publications
• Sorted by Date • Classified by Publication Type • Classified by Research Category •
Michael Beetz, Freek Stulp, Alexandra Kirsch, Armin Müller, and Sebastian Buck. Autonomous Robot Controllers Capable of Acquiring Repertoires of Complex Skills. In RoboCup-2003: The Seventh RoboCup Competitions and Conferences, Springer Verlag, 2003.
Download
Abstract
Due to the complexity and sophistication of the skills needed in real world tasks, the development of autonomous robot controllers requires an ever increasing application of learning techniques. To date, however, learning steps are mainly executed in isolation and only the learned code pieces become part of the controller. This approach has several drawbacks: the learning steps themselves are undocumented and not executable. In this paper, we extend an existing control language with constructs for specifying control tasks, process models, learning problems, exploration strategies, etc. Using these constructs, the learning problems can be represented explicitly and transparently and, as they are part of the overall program implementation, become executable. With the extended language we rationally reconstruct large parts of the action selection module of the Agilo2001 autonomous soccer robots.
BibTeX Entry
@InProceedings{beetz03autonomous,
author = {Michael Beetz and Freek Stulp and Alexandra Kirsch and Armin M\"uller and Sebastian Buck},
title = {Autonomous Robot Controllers Capable of Acquiring Repertoires of Complex Skills},
booktitle = {{R}obo{C}up-2003: The Seventh {R}obo{C}up Competitions and Conferences},
publisher = {Springer Verlag},
year = {2003},
editor = {Daniel Polani and Andrea Bonarini and Brett Browning and Kazuo Yoshida},
pdf = {http://www9.in.tum.de/publications/2003/ROBOCUP-SYMP-03-Beetz-etal.pdf},
url = {http://citeseer.ist.psu.edu/beetz03autonomous.html},
bib2html_pubtype = {Refereed Conference Paper},
bib2html_rescat = {Robotics - Motor Control},
abstract = {Due to the complexity and sophistication of the skills
needed in real world tasks, the development of autonomous robot controllers
requires an ever increasing application of learning techniques. To date,
however, learning steps are mainly executed in isolation and only the learned
code pieces become part of the controller. This approach has several
drawbacks: the learning steps themselves are undocumented and not executable.
In this paper, we extend an existing control language with constructs for
specifying control tasks, process models, learning problems, exploration
strategies, etc. Using these constructs, the learning problems can be
represented explicitly and transparently and, as they are part of the overall
program implementation, become executable. With the extended language we
rationally reconstruct large parts of the action selection module of the
Agilo2001 autonomous soccer robots.}
}
Generated by bib2html.pl (written by Patrick Riley ) on Tue Sep 23, 2008 15:09:11 UTC

















