MULTIPLE MODEL ADAPTIVE CONTROL FOR INTERVAL SYSTE... Copyright © 1999 IFAC 14th Triennial \Vorld P.R. China MULTIPLE MODEL ADAPTIVE CONTROL FOR INTERVAL SYSTE,MS Youngchol Kim, Sunwook Choi'l Taeshin Cho 14th World Congress of IFAC 1-3b-09-4 School ofElectrical and Electronics Eng., JvTational [lniversily' San 48, Gaesindong" Cheongju. Ch ungbuk, 361-763, Korea E-mail: Abstract : We present a multiple rnodel adaptiye control(l\11v1AC:) for linear interval systelns of real coefficient_ In order to improye the perfonnance as ,veIl as stability, the multiple model adaptiYe control(1VII\t1AC) method is introduced. In tlris approac4 the identification models are paranlcterized in a finite set \,lLich consist s in 16 Kharitono'" systems and each controIIer of multiple models is designed by the coefficient diagram method(CDM) ,'ith PlD structure. TIle characteristic equation of overall systenl results in a set of nlultilincar perturbed polynolnials \,ith respect to uncertain pardmeters. The robust stability of the proposed method are analyzed and the perfonnance is examined by SilTIulations on example. Copyright © 1999 IFAC Key,Yords : Kharitonov Interval system, CDM 1. INTRODUCTION \\lhen ,ye consider a robust design problerll for uncertain plants it is desirable to design a control system ,,"hich is not only stable but also guarantees an adequate level of perfonnance. iVlnltiple Model Adapti"e 1995) is such a Inethodo log} that can leanl to operate efficiently in dynamical enyiromnents possessing a high degree of unccrtainl}. Multiple IIlOdcls arc used to describe the different eoyironments and the control is effected by a linear cOlubination of controllers determined different models. The MJ\1AC is not lle\y in control theory. In fact multiple Kalman filters ,,"ere proposed in the 1970s by This ,,·ork \"as supported in part by MOST of under 97 ... T-Ol-A-lOl Copyright 1999 IFAC Lainiotis and i\th£JnS et al.(Alhans et J979) to improye the accuracy of the state estilnate in control problellls. Iv1any interesting practical applications of this approach haye appeared in the follo\ving t\yO decades. Kaufman and colleagues (He, et aI,,> 1986) applied it to Inedical ,Ybile Moose and co- \yorkers used these ideas in meneuverable target tracking. Application in air traffic control \yas proposed by BarShalom and and Maybaek and et 1989) adapted it to fault detection and control in aircraft. all this cases haye not reported stability results for these systems. s\vitching schemes for this approach haye appear in the literature(He, et al., T'1is case should haye a lot of nlodels to satisfy perfonnance and the stability by s\vitching. This paper deals "ith tile follo\ving objectives: The first is to extend to interval system \vith ISBN: 0 08 043248 4 4699