Force Control by RBF Fuzzy Neuro with Unsupervised Learning

Recently, Fuzzy reasoning has been used in many fields of applications. In the application of the fuzzy reasoning to various fields, the tuning and optimizing method is the key issue. Some self-tuning methods have been proposed so far. Most of these conventional self-tuning methods require teaching...

Full description

Saved in:
Bibliographic Details
Published inTransactions of the Japan Society of Mechanical Engineers Series C Vol. 61; no. 588; pp. 3311 - 3317
Main Authors Shimojima, Koji, Fukuda, Toshio, Hasegawa, Yasuhisa
Format Journal Article
LanguageJapanese
Published The Japan Society of Mechanical Engineers 25.08.1995
Subjects
Feedforward Controller
Fuzzy Controller
Genetic Algorithm
Radial Basis Function
Unsupervised Learning
Online AccessGet full text
ISSN0387-5024
1884-8354
DOI10.1299/kikaic.61.3311

Cover

More Information
Summary:Recently, Fuzzy reasoning has been used in many fields of applications. In the application of the fuzzy reasoning to various fields, the tuning and optimizing method is the key issue. Some self-tuning methods have been proposed so far. Most of these conventional self-tuning methods require teaching data or a mathematical model of the environment in order to tune the fuzzy systems. However teaching data are not obtained in many cases, because of the nonlinear property of the control objects. In this paper, we propose a new unsupervised self-tuning method for fuzzy resoning. The fuzzy reasoning consists of some membership functions expressed by the radial basis function with an insensitive region. The unsupervised learning is carried out by the genetic algorithm, which has two mutation operators for global and local searches. The effectiveness of the proposed method is shown through simulations of the force control for uncertain objects.
ISSN:0387-5024
1884-8354
DOI:10.1299/kikaic.61.3311