Creating a CMAC with overlapping basis functions in order to prevent weight drift

Creating a CMAC with overlapping basis functions in order to prevent weight drift The cerebellar model articulation controller, or CMAC, is a type of associative memory neural network suitable for use in direct adaptive control schemes. However, the CMAC exhibits a large trade-off between stability and performance when inputs oscillate. This is due to the local nature of the basis functions—an input oscillating between two basis functions on one layer can cause their weights to drift in opposite directions. Continued drift will eventually affect performance, resulting in bursting. The proposed method overlaps the basis functions on each layer so that an oscillation will occur within basis functions. This makes the weights much less prone to drift. A simulation with a flexible joint demonstrates that both high performance and stability can be achieved using the proposed method. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Soft Computing Springer Journals

Creating a CMAC with overlapping basis functions in order to prevent weight drift

Soft Computing , Volume 21 (16) – May 27, 2016
Loading next page...
 
/lp/springer_journal/creating-a-cmac-with-overlapping-basis-functions-in-order-to-prevent-iWQl6N0NLS
Publisher
Springer Berlin Heidelberg
Copyright
Copyright © 2016 by Springer-Verlag Berlin Heidelberg
Subject
Engineering; Computational Intelligence; Artificial Intelligence (incl. Robotics); Mathematical Logic and Foundations; Control, Robotics, Mechatronics
ISSN
1432-7643
eISSN
1433-7479
D.O.I.
10.1007/s00500-016-2204-0
Publisher site
See Article on Publisher Site

Abstract

The cerebellar model articulation controller, or CMAC, is a type of associative memory neural network suitable for use in direct adaptive control schemes. However, the CMAC exhibits a large trade-off between stability and performance when inputs oscillate. This is due to the local nature of the basis functions—an input oscillating between two basis functions on one layer can cause their weights to drift in opposite directions. Continued drift will eventually affect performance, resulting in bursting. The proposed method overlaps the basis functions on each layer so that an oscillation will occur within basis functions. This makes the weights much less prone to drift. A simulation with a flexible joint demonstrates that both high performance and stability can be achieved using the proposed method.

Journal

Soft ComputingSpringer Journals

Published: May 27, 2016

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off