Vol.6, Special Issue, 2007 pp. 25-38
UDC 681.5.01:007.52(045)=111
ROBUST ADAPTIVE CONTROL OF MECHATRONIC
SYSTEMS EMPLOYING AN EMULATION OF NONLINEAR FUNCTIONS
Yuan-Wei Jing1, Yan-Xin Zhang1,
Tatjana Kolemeisevska-Gugulovska2, Georgi
M. Dimirovski2,3, Miomir K. Vukobratovic4
1Northeastern University, School of
Information Sciences & Engineering Shenyang,110004, Liaoning, P. R.
of China
2SS Cyril & Methodius University,
Faculty of Electrical Eng. & Info. Technologies Rugjer Boskovic BB,
Karpos II, 1000, Skopje, Republic of Macedonia
3Dogus University, Faculty of Engineering,
Dept. of Computer Engineering Acibadem, Zeamet Sk. 21, Kadikoy, 34722 Istanbul,
Republic of Turkey
4Mıhailo Pupin Institute, Laboratory
of Robotics & Automation Volgina 15, Karaburma, 11000, Republic of
Serbia
e-mail: ywjjing@ mail.neu.edu.cn;
yanxin@ 126.com; tanjakg@
feit.ukim.edu.mk; gdimirovski@
dogus.edu.tr; vuk@ robot.imp.bg.ac.yu
Abstract. A novel robust adaptive control design synthesis, which
employs both high-order neural networks and math-analytical results, for
a class of mechatronic nonlinear systems possessing similarity property
has been derived. This approach makes an adequate use of the structural
feature of composite similarity systems and neural networks to resolve
the representation issue of uncertainty interconnections and subsystem
gains by on-line updating the weights. This synthesis does guarantee the
real stability in closed-loop but requires skills to obtain larger attraction
domains. Mechatronic example of an axis-tray drive system, possessing uncertainties,
is used to illustrate the proposed technique.
Key words: Adaptive control, function emulation, mechatronic
systems, neural networks, stability synthesis
ROBUSTNO ADAPTIVNO UPRAVLJANJE MEHATRONIČNIH
SISTEMA KOJI KORISTE
ANN EMULACIJU NONLINEARNIH FUNKCIJA
Izvodi se nova sinteza robustnog adaptivnog upravljanja koje koristi kako
neuronske mreže visokog rede tako i matematičko-analitičke rezultate
za klasu mehatroničkih nonlinearnih sistema sa sličnostima u ovom radu.
Ovaj pristup predstavlja adekvatnu upotrebu strukturalnih osobina složenih
sistema sa sličnostima i neuronskih mreža da bi razrešile neizvesno
predstavljanje interkonekcija i pojačanja podsistema onlajn obnavljanjem
težina. Ova sinteza garantuje realnu stabilnost u zatvorenim petljama
i zahteva projektantsku veštinu da bi se postigli veći domeni privljačenja.
Koristi se mehatronički primer prenosno-spregajuceg sistema ose, koji
ima neizvesnosti, da bi se ilustrovala predložena tehnika.
Ključne reči: Adaptivna upravljanje, emulacija funkcija, mehatronički
sistemi, neronske mreže, sinteza stabilnosti
