Legs that can walk: Embodiment-Based Modular Reinforcement Learning applied

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Citations (Scopus)
141 Downloads (Pure)

Abstract

Experiments to illustrate a novel methodology for
reinforcement learning in embodied physical agents are described.
A simulated legged robot is decomposed into structurebased
modules following the authors' EMBER principles of
local sensing, action and learning. The legs are individually
trained to 'walk' in isolation, and re-attached to the robot;
walking is then sufficiently stable that learning in situ can
continue. The experiments demonstrate the benefits of the
modular decomposition: state-space factorisation leads to faster
learning, in this case to the extent that an otherwise intractable
problem becomes learnable.
Original languageEnglish
Title of host publicationProcs 2005 IEEE Int Symposium on Computational Intelligence in Robotics and Automation
Subtitle of host publicationCIRA 2005
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages365-372
ISBN (Print)0-7803-9355-4
Publication statusPublished - 2005
Event2005 IEEE Int Symposium of Computational Intelligence in Robotics & Automation - Espoo, Finland
Duration: 27 Jun 200530 Jun 2005

Conference

Conference2005 IEEE Int Symposium of Computational Intelligence in Robotics & Automation
Country/TerritoryFinland
CityEspoo
Period27/06/0530/06/05

Fingerprint

Dive into the research topics of 'Legs that can walk: Embodiment-Based Modular Reinforcement Learning applied'. Together they form a unique fingerprint.

Cite this