Humanoids Learning to Walk: a Natural CPG-Actor-Critic Architecture

Humanoids Learning to Walk: a Natural CPG-Actor-Critic Architecture

The identification of learning mechanisms for locomotion has been the subject of much researchfor some time but many challenges remain. Dynamic systems theory (DST) offers a novel approach to humanoid learning through environmental interaction. Reinforcement learning (RL) has offered a promising me...

Full description

Bibliographic Details
Main Authors: CAI eLI, Robert eLowe, Tom eZiemke
Format: Article
Language:English
Published: Frontiers Media S.A. 2013-04-01
Series:Frontiers in Neurorobotics
Subjects:
reinforcement learning
SI
cpg-actor-critic
DST
value system
Online Access:http://journal.frontiersin.org/Journal/10.3389/fnbot.2013.00005/full

Similar Items

Similar Items