Gain scheduling control of nonlinear plant using RBF neural network

Abstract

In this paper, an on-line approach to gain scheduling control of a type of nonlinear plant is proposed. The method consists of a partitioning algorithm to partition the plant's operating space into several regions, a mechanism that designs a linear controller for each region, and a radial basis function neural network (RBFN) for on-line interpolation of the controller parameters among the different regions. The method is described in detail, and is studied analytically in computer simulation on gain scheduled PI control of a nonlinear plant, which shows encouraging performance.