by Michael Beetz, Alexandra Kirsch and Armin Müller
Abstract:
In this paper, we extend the autonomous robot control and plan language RPL with constructs for specifying experiences, control tasks, learning systems and their parameterization, and exploration strategies. Using these constructs, the learning problems can be represented explicitly and transparently and become executable. With the extended language we rationally reconstruct parts of the AGILO autonomous robot soccer controllers and show the feasibility and advantages of our approach.
Reference:
Michael Beetz, Alexandra Kirsch and Armin Müller, "RPL-LEARN: Extending an Autonomous Robot Control Language to Perform Experience-based Learning", In 3rd International Joint Conference on Autonomous Agents & Multi Agent Systems (AAMAS), 2004.
Bibtex Entry:
@InProceedings{Bee04RPL,
author = {Michael Beetz and Alexandra Kirsch and Armin M{\"u}ller},
title = {{RPL-LEARN}: Extending an Autonomous Robot Control Language to Perform Experience-based Learning},
booktitle = {3rd International Joint Conference on Autonomous Agents \& Multi Agent Systems (AAMAS)},
year = {2004},
bib2html_pubtype = {Conference Paper},
bib2html_rescat = {Learning},
bib2html_groups = {AGILO,Cogito},
bib2html_funding = {AGILO},
bib2html_keywords = {Learning, Robot, Language},
abstract = {In this paper, we extend the autonomous robot control and plan language RPL with constructs for
specifying experiences, control tasks, learning systems and their parameterization, and exploration
strategies. Using these constructs, the learning problems can be represented explicitly and
transparently and become executable. With the extended language we rationally reconstruct parts of
the AGILO autonomous robot soccer controllers and show the feasibility and advantages of our
approach.}
}