SINGLE NET INDIRECT LEARNING ARCHITECTURE

被引：12

作者：

ANDERSEN, HC

TENG, FC

TSOI, AC

机构：

[1] Department of Electrical and Computer Engineering, University of Queensland, Brisbane

来源：

IEEE TRANSACTIONS ON NEURAL NETWORKS | 1994年 / 5卷 / 06期

基金：

澳大利亚研究理事会;

关键词：

Controller network training - Neural controller - Neurons - Single net indirect learning architecture;

D O I：

10.1109/72.329701

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This letter presents a novel indirect learning architecture which uses a single neural network in implementation. The new architecture generates the error signal required in training the controller network by an innovative design using a memory element and few switches. The new controller needs only half the number of neurons and connection weights in comparison with the original indirect learning architecture. Also given are the simulation results in controlling a nonlinear plant.

引用

页码：1003 / 1005

页数：3

共 8 条

[1]

[Anonymous], INTRO THEORY NEURAL

[2]

Chen F.-C., 1990, IEEE Control Systems Magazine, V10, P44, DOI 10.1109/37.55123

[3]

GARCIAPADILLA F, 1991, P INT WORKSHOP IWANN

[4]

JORDAN MI, 1992, IN PRESS COGNITIVE S

[5]

LAN MS, 1989, P AUTO CONTR C, P910

[6]

Psaltis D., 1988, IEEE Control Systems Magazine, V8, P17, DOI 10.1109/37.1868

[7]

TENG FC, 1992, IEEE T NEURAL NETS

[8] AN ADAPTIVE-CONTROL FOR CARMA SYSTEMS USING LINEAR NEURAL NETWORKS [J].

WATANABE, K ;

FUKUDA, T ;

TZAFESTAS, SG .

INTERNATIONAL JOURNAL OF CONTROL, 1992, 56 (02) :483-497

← 1 →