Informatics Report Series


Report   

EDI-INF-RR-0948


Related Pages

Report (by Number) Index
Report (by Date) Index
Author Index
Institute Index

Home
Title:Value Function Approximation on Non-Linear Manifolds for Robot Motor Control
Authors: Masashi Sugiyama ; Hirotaka Hachiya ; Christopher Towell ; Sethu Vijayakumar
Date:Apr 2007
Publication Title:Robotics and Automation, 2007 IEEE International Conference on
Publisher:IEEE
Publication Type:Conference Paper Publication Status:Published
Page Nos:1733-1740
DOI:10.1109/ROBOT.2007.363573 ISBN/ISSN:1-4244-0601-3
Abstract:
The least squares approach works efficiently in value function approximation, given appropriate basis functions. Because of its smoothness, the Gaussian kernel is a popular and useful choice as a basis function. However, it does not allow for discontinuity which typically arises in realworld reinforcement learning tasks. In this paper, we propose a new basis function based on geodesic Gaussian kernels, which exploits the non-linear manifold structure induced by the Markov decision processes. The usefulness of the proposed method is successfully demonstrated in a simulated robot arm control and Khepera robot navigation.
Copyright:
2007 by The University of Edinburgh. All Rights Reserved
Links To Paper
No links available
Bibtex format
@InProceedings{EDI-INF-RR-0948,
author = { Masashi Sugiyama and Hirotaka Hachiya and Christopher Towell and Sethu Vijayakumar },
title = {Value Function Approximation on Non-Linear Manifolds for Robot Motor Control},
book title = {Robotics and Automation, 2007 IEEE International Conference on},
publisher = {IEEE},
year = 2007,
month = {Apr},
pages = {1733-1740},
doi = {10.1109/ROBOT.2007.363573},
}


Home : Publications : Report 

Please mail <reports@inf.ed.ac.uk> with any changes or corrections.
Unless explicitly stated otherwise, all material is copyright The University of Edinburgh