A linear approach to generalized minimum variance controller design for MIMO nonlinear systems

作者：Alipouri Yousef^*; Poshtan Javad

来源：Nonlinear Dynamics, 2014, 77(3): 935-949.

DOI：10.1007/s11071-014-1352-y

• Generalized minimum variance controller
• VARX model
• MIMO nonlinear system
• Discrete Taylor series
• Four-tank benchmark system
• Interactor matrix method

Designing minimum variance controllers (MVC) for nonlinear systems is confronted with many difficulties. The methods which are able to identify MIMO nonlinear systems are scarce, and linear models are not accurate in modeling nonlinear systems. In this paper, Vector ARX (VARX) models are proposed for designing MVC and generalized minimum variance controller (GMVC) for linear and nonlinear systems, and the accuracy of these models in approximating the nonlinear MIMO system is studied. However, the VARX is a linear model. It is shown that this model can identify some kinds of nonlinear systems with any desired accuracy. Therefore, the controller designed by the VARX is accurate, even for these nonlinear systems. The proposed controller is tested on a both linear system and a nonlinear four-tank benchmark process. In spite of the simplicity of designing GMVCs for the VARX models, the results show that the proposed method is accurate and implementable.

• 出版日期2014-8