| Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 90 === Abstract
In this thesis, we studied the control problem of using adaptive fuzzy controllers for nonlinear unstable systems. The model of this controller is the Takagi and Sugeno fuzzy model. In this thesis, how the parameter matrix P required in the Lyapunov equation for guaranteeing the system stability can affect the system performance are discussed. It can be found that even though the Lyapunov synthesis approach has already proven the system stability, the parameter matrix P must still be selected appropriately beside of the symmetric positive definite property. In our simulations, it can also be found that this kind of controller does not converge to a fixed controller, but it adaptively adjusts its parameters according the errors. As a consequence, when the considered system has sensory noise on the states, the system may gradually become unstable. Two way of restraining such an unbounded phenomenon are proposed. One is to use a bound constraint directly for the parameters in the consequences of fuzzy rules. The other is to use a bound constraint for the norm of all parameters. Both approaches can have nice effects to restrain the unbounded signals. From our simulations, we found that the latter one has the better performances.
|
|---|
