Research On The Generalized Predictive Control Of A Class Of Nonlinear System

Generalized Predictive Control(GPC) for a class of nonlinear system is studied in this paper. The main results are concluded as follows:(1) The method that a class of nonlinear can be substituted by a time-varying linear system is given and proved. So theories of linear systems can be used for nonlinear directly. Least square identification and orthogonal polynomials are used for the time-varying parameters identification . An implicit GPC for nonlinear system is presented, in which orthogonal polynomials are used to estimated the time-varying parameters of the generalized predictive controller. The stability of the closed-loop system is analyzed.(2) Wavelet networks are introduced to GPC. Several nonlinear GPC algorithms based on wavelet networks are given: a nonlinear GPC base on predictive error compensation is designed , in which wavelet network is used to model the predictive error; A new structure of multi-step predictive controller is constructed in which wavelet networks are used to identified the nonlinear directly; A single-step predictive controller is given, in which a wavelet networks are used to estimate the parameters of the linear time-varying system.; An a implicit GPC for nonlinear system is presented, in which wavelet networks are used to estimated the time-varying parameters of the generalized predictive controller. For the above methods, the performance of the closed-loop system is analyzed(3) Nonlinear GPC based on wavelet networks inverse system is proposed, in which wavelet network is used to model the a - inverse system of the nonlinear system. The structure of internal control is adapted for GPC. The stability and tracking performance with modeling error are discussed.(4) The equivalent linearization is extended to nonlinear multivariable system. An generalized predictive controllers for the nonlinear system is listed . Decouple GPC for multivariable nonlinear is studied. A new decouple GPC strategy is designed, in which wavelet networks are used to model the generalized inverse model of the nonlinear system.(5) The idea of the equivalent linearization is extended to a class of nonlinear continuous system. The continuous nonlinear generalized predictive controllers for the nonlinear system are designed.