Rho-learning: a robotics oriented reinforcement learning algorithm

We present a new reinforcement learning system more suitable to be used in robotics than existing ones. Existing reinforcement learning algorithms are not specifically tailored for robotics and so they do not take advantage of the robotic perception characteristics as well as of the expected complex...

ver descrição completa

Detalhes bibliográficos
Autor: Porta, Josep M.
Tipo de documento: outro
Data de publicação:2000
País:España
Recursos:Consejo Superior de Investigaciones Científicas (CSIC)
Repositório:DIGITAL.CSIC. Repositorio Institucional del CSIC
OAI Identifier:oai:digital.csic.es:10261/29985
Acesso em linha:http://hdl.handle.net/10261/29985
Access Level:Acceso aberto
Palavra-chave:Reinforcement learning
Robot learning
Sensor relevance
Walking robots
Automation
Robots
id ES_5fe32cdd7b9d30aa25d572e9d6c29fa7
oai_identifier_str oai:digital.csic.es:10261/29985
network_acronym_str ES
network_name_str España
repository_id_str
spelling Rho-learning: a robotics oriented reinforcement learning algorithmPorta, Josep M.Reinforcement learningRobot learningSensor relevanceWalking robotsAutomationRobotsWe present a new reinforcement learning system more suitable to be used in robotics than existing ones. Existing reinforcement learning algorithms are not specifically tailored for robotics and so they do not take advantage of the robotic perception characteristics as well as of the expected complexity of the task that robots are likely to face. In a robot, the information about the environment comes from a set of qualitatively different sensors and in the main par of tasks small subsets of these sensors provide enough information to correctly predict the effect of actions. Departing from this analysis, we outline a new reinforcement learning system that aims at determining relevant subsets of sensors for each action and we present an algorithm that partially implements this new reinforcement learning architecture. Results of the application of the algorithm to the problem of learning to walk with a six legged robot are presented and compared with a well known reinforcement learning algorithm (Q-learning) showing the advantages of our approach.201020102000info:eu-repo/semantics/otherhttp://purl.org/coar/resource_type/c_18ghinfo:eu-repo/semantics/reporthttp://hdl.handle.net/10261/29985reponame:DIGITAL.CSIC. Repositorio Institucional del CSICinstname:Consejo Superior de Investigaciones Científicas (CSIC)Inglésinfo:eu-repo/semantics/openAccessoai:digital.csic.es:10261/299852026-05-22T06:33:51Z
dc.title.none.fl_str_mv Rho-learning: a robotics oriented reinforcement learning algorithm
title Rho-learning: a robotics oriented reinforcement learning algorithm
spellingShingle Rho-learning: a robotics oriented reinforcement learning algorithm
Porta, Josep M.
Reinforcement learning
Robot learning
Sensor relevance
Walking robots
Automation
Robots
title_short Rho-learning: a robotics oriented reinforcement learning algorithm
title_full Rho-learning: a robotics oriented reinforcement learning algorithm
title_fullStr Rho-learning: a robotics oriented reinforcement learning algorithm
title_full_unstemmed Rho-learning: a robotics oriented reinforcement learning algorithm
title_sort Rho-learning: a robotics oriented reinforcement learning algorithm
dc.creator.none.fl_str_mv Porta, Josep M.
author Porta, Josep M.
author_facet Porta, Josep M.
author_role author
dc.subject.none.fl_str_mv Reinforcement learning
Robot learning
Sensor relevance
Walking robots
Automation
Robots
topic Reinforcement learning
Robot learning
Sensor relevance
Walking robots
Automation
Robots
description We present a new reinforcement learning system more suitable to be used in robotics than existing ones. Existing reinforcement learning algorithms are not specifically tailored for robotics and so they do not take advantage of the robotic perception characteristics as well as of the expected complexity of the task that robots are likely to face. In a robot, the information about the environment comes from a set of qualitatively different sensors and in the main par of tasks small subsets of these sensors provide enough information to correctly predict the effect of actions. Departing from this analysis, we outline a new reinforcement learning system that aims at determining relevant subsets of sensors for each action and we present an algorithm that partially implements this new reinforcement learning architecture. Results of the application of the algorithm to the problem of learning to walk with a six legged robot are presented and compared with a well known reinforcement learning algorithm (Q-learning) showing the advantages of our approach.
publishDate 2000
dc.date.none.fl_str_mv 2000
2010
2010
dc.type.none.fl_str_mv info:eu-repo/semantics/other
http://purl.org/coar/resource_type/c_18gh
dc.type.openaire.fl_str_mv info:eu-repo/semantics/report
format other
dc.identifier.none.fl_str_mv http://hdl.handle.net/10261/29985
url http://hdl.handle.net/10261/29985
dc.language.none.fl_str_mv Inglés
language_invalid_str_mv Inglés
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.source.none.fl_str_mv reponame:DIGITAL.CSIC. Repositorio Institucional del CSIC
instname:Consejo Superior de Investigaciones Científicas (CSIC)
instname_str Consejo Superior de Investigaciones Científicas (CSIC)
reponame_str DIGITAL.CSIC. Repositorio Institucional del CSIC
collection DIGITAL.CSIC. Repositorio Institucional del CSIC
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869409251207675904
score 15,812455

Registros relacionados