ROBUST MULTIPLE MODEL ADAPTIVE CONTROL WITH FUZZY POSTERIOR PROBABILITY COMBINATION

Fatemeh Zare-Mirakabad | Mohammad H. Kazemi

---

URL :   http://research.shahed.ac.ir/WSR/WebPages/Report/PaperView.aspx?PaperID=43547
Date :  2017/04/04
Publish in :    Control and Intelligent Systems-It,s Title Changed
DOI :  https://doi.org/10.2316/journal.201.2017.2.201-2796
Link :  http://www.actapress.com/PaperInfo.aspx?paperId=45598
Keywords :CONTROL, POSTERIOR

Abstract :

---

This paper proposes a fuzzy posterior probability (FPP) combina- tion in a robust multiple model adaptive control (MMAC) frame- work, for linear systems subject to uncertain real parameters, either constant or slowly time varying. The proposed control scheme sub- stituted posterior probability evaluator in robust MMAC architec- ture for FPP. A Takagi–Sugeno–Kang-type fuzzy rule-based system is proposed to generate the weights for probabilistic weighting of the local controls to form the global signal control. The local controls are designed using robust mixed-µ synthesis of the plant evaluated in a set of predefined values of uncertain parameters so that the local stability and performance robustness are guaranteed. Local Kalman filters are also designed to produce residual signals which are utilized by the FPP. The proposed scheme is applied to controller synthesis of a two-cart mass–spring–damper system. The simulation results illustrate the advantages of the proposed FPP against its conventional evaluator.