An improvement of quasi-ARX predictor to control of nonlinear systems using nonlinear PCA network

In this paper, a nonlinear principal component analysis (NPCA) is introduced to improve the quasi-ARX modeling. One part of the quasi-ARX model is an ordinary neurofuzzy network to parameterize the coefficients which faces to a problem of high dimension. NPCA is used for this part to deal with the problem. The processes of modeling, parameter estimating and control are given detailedly. Some simulations of systems controlling are provided to illustrate the effectiveness of the proposed modeling approach.