Adaptive Inverse Control Based on Volterra Polynomial Basis Function Neural Networks
Dang Yingnong, Han Chongzhao
(Xi'an Jiaotong University, Xi'an 710049, China)

Abstract: An adaptive inverse control method for nonlinear systems is proposed by using the Volterra polynomial basis function (VPBF) neural networks. VPBF neural networks were used to describe both the plant and the inverse controller. The structure and initial parameters of the VPBF neural networks were determined by using the orthogonal leastª²squares algorithm. A dynamical normalized nonlinear least mean square (DNNLMS) algorithm was developed for onª²line weights learning. The proposed method is simple and has fast learning rate.
Keywords: nonlinear systems control;Volterra polynomial basis function;adaptive inverse control