Reference governor control of constrained feedback systems using neural networks

Author

Jahagirdar, Harshad ; Keerthi, S. Sathiya ; Ang, M.H., Jr.

Author_Institution

Dept. of Mech. Eng., Nat. Univ. of Singapore, Singapore

fYear

2002

fDate

2002

Firstpage

223

Lastpage

227

Abstract

A neural network approach to reference governor control of systems with constraints on state and control variables is discussed. A feed-forward neural network architecture is used to define safety sets in a constrained system state-space. Results presented here include the description of a neural reference governor algorithm and its application to linear and nonlinear control systems. The objective is to demonstrate the feasibility of such a design as an alternative to the Lyapunov function approach to the control of constrained systems.

Keywords

backpropagation; control system synthesis; feedback; feedforward neural nets; linear systems; neurocontrollers; nonlinear control systems; state-space methods; constrained feedback systems; constrained system state-space; control variables; feedforward neural network architecture; linear control systems; neural networks; neural reference governor algorithm; nonlinear control systems; reference governor control; safety sets; state variables; two-layer feed-forward backpropagation neural network; Automatic control; Control systems; Equations; Mechanical variables control; Neural networks; Neurofeedback; Optimal control; Safety; Signal design; State feedback;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control, 2002. Proceedings of the 2002 IEEE International Symposium on

ISSN

2158-9860

Print_ISBN

0-7803-7620-X

Type

conf

DOI

10.1109/ISIC.2002.1157766

Filename

1157766