Learning multiple fuzzy control for robot manipulators

Keigo Watanabe, Sang Ho Jin, Spyros G. Tzafestas

Research output: Contribution to journalArticle

Abstract

An iterative learning fuzzy controller is proposed for controlling the trajectory of a robot manipulator. The learning controller described here is called a multiple fuzzy controller, in which several elemental fuzzy controllers are processed in parallel, and the degree of usage of each inferred consequent is determined by using a linear neural network. Two learning algorithms are examined in the framework of the specialized learning architecture: one is based on minimizing the squared sum of the trajectory error for each link, and the other is based on directly minimizing the squared trajectory error for each link. The effectiveness of the present control method is illustrated by using some numerical simulations for a two-link manipulator.

Original languageEnglish
Pages (from-to)119-136
Number of pages18
JournalJournal of artificial neural networks
Volume2
Issue number1-2
Publication statusPublished - 1995
Externally publishedYes

Fingerprint

Fuzzy control
Manipulators
Robots
Controllers
Trajectories
Learning algorithms
Neural networks
Computer simulation

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Learning multiple fuzzy control for robot manipulators. / Watanabe, Keigo; Jin, Sang Ho; Tzafestas, Spyros G.

In: Journal of artificial neural networks, Vol. 2, No. 1-2, 1995, p. 119-136.

Research output: Contribution to journalArticle

Watanabe, Keigo ; Jin, Sang Ho ; Tzafestas, Spyros G. / Learning multiple fuzzy control for robot manipulators. In: Journal of artificial neural networks. 1995 ; Vol. 2, No. 1-2. pp. 119-136.
@article{fc8c2006a21e4a338e9f7f4b4e7f2f52,
title = "Learning multiple fuzzy control for robot manipulators",
abstract = "An iterative learning fuzzy controller is proposed for controlling the trajectory of a robot manipulator. The learning controller described here is called a multiple fuzzy controller, in which several elemental fuzzy controllers are processed in parallel, and the degree of usage of each inferred consequent is determined by using a linear neural network. Two learning algorithms are examined in the framework of the specialized learning architecture: one is based on minimizing the squared sum of the trajectory error for each link, and the other is based on directly minimizing the squared trajectory error for each link. The effectiveness of the present control method is illustrated by using some numerical simulations for a two-link manipulator.",
author = "Keigo Watanabe and Jin, {Sang Ho} and Tzafestas, {Spyros G.}",
year = "1995",
language = "English",
volume = "2",
pages = "119--136",
journal = "Journal of artificial neural networks",
issn = "1073-5828",
number = "1-2",

}

TY - JOUR

T1 - Learning multiple fuzzy control for robot manipulators

AU - Watanabe, Keigo

AU - Jin, Sang Ho

AU - Tzafestas, Spyros G.

PY - 1995

Y1 - 1995

N2 - An iterative learning fuzzy controller is proposed for controlling the trajectory of a robot manipulator. The learning controller described here is called a multiple fuzzy controller, in which several elemental fuzzy controllers are processed in parallel, and the degree of usage of each inferred consequent is determined by using a linear neural network. Two learning algorithms are examined in the framework of the specialized learning architecture: one is based on minimizing the squared sum of the trajectory error for each link, and the other is based on directly minimizing the squared trajectory error for each link. The effectiveness of the present control method is illustrated by using some numerical simulations for a two-link manipulator.

AB - An iterative learning fuzzy controller is proposed for controlling the trajectory of a robot manipulator. The learning controller described here is called a multiple fuzzy controller, in which several elemental fuzzy controllers are processed in parallel, and the degree of usage of each inferred consequent is determined by using a linear neural network. Two learning algorithms are examined in the framework of the specialized learning architecture: one is based on minimizing the squared sum of the trajectory error for each link, and the other is based on directly minimizing the squared trajectory error for each link. The effectiveness of the present control method is illustrated by using some numerical simulations for a two-link manipulator.

UR - http://www.scopus.com/inward/record.url?scp=0029484412&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029484412&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0029484412

VL - 2

SP - 119

EP - 136

JO - Journal of artificial neural networks

JF - Journal of artificial neural networks

SN - 1073-5828

IS - 1-2

ER -