DocumentCode :
2815961
Title :
Adaptive control systems switched by control and robust performance criteria
Author :
Inaba, Masaaki ; Guo, Hai-Jiao ; Nakao, Kazuo ; Abe, Kenichi
Author_Institution :
Dept. of Electr. Eng., Tohoku Univ., Sendai, Japan
Volume :
2
fYear :
1996
fDate :
18-21 Nov 1996
Firstpage :
690
Abstract :
This paper presents a methodology for the design of adaptive control systems using multiple controllers which can control an unknown plant whose parameters change widely. It is not necessary to estimate the plant parameter values in our system. Multiple controllers are constructed by several basic controllers. Switching from the current basic controller to a more appropriate one is based on one-way simple switching rules. The switching rules only utilize the information of whether the control has succeeded or failed. We introduce the control and robust performance criteria. The former is used to produce a number of candidates for the basic controllers for typical plant models. The latter is used to select several basic controllers from the candidates to construct multiple controllers in order to achieve the system robust performance. To illustrate the effectiveness of the approach, we apply the control system to the cart-pole problem with changeable length of the pole
Keywords :
adaptive control; control system synthesis; genetic algorithms; intelligent control; learning systems; robust control; adaptive control systems; cart-pole problem; genetic algorithm; learning process; multiple controllers; robust performance criteria; switching rules; Adaptive control; Control systems; Degradation; Design methodology; Fuzzy control; Noise measurement; Parameter estimation; Programmable control; Robots; Robust control;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Emerging Technologies and Factory Automation, 1996. EFTA '96. Proceedings., 1996 IEEE Conference on
Conference_Location :
Kauai, HI
Print_ISBN :
0-7803-3685-2
Type :
conf
DOI :
10.1109/ETFA.1996.573988
Filename :
573988
Link To Document :
بازگشت