Abstract��A modeling method for nonlinear dynamic system based on Support Vector Regression (SVR) was proposed in this paper. The Hammerstein model expressed by a nonlinear static subunit followed by a linear dynamic subunit was used to describe the nonlinear dynamic system. Through the function expansion, the nonlinear transfer function of Hammerstein model could be converted to the same form as linear one, thus generating the intermediate linear model. Also, by SVR algorithm, the coefficients of the intermediate model were obtained. Moreover, through the relations of the coefficients of intermediate model and that of Hammerstein model, the nonlinear static subunit and linear dynamic subunit were identified simultaneously. Calibrating experimental data of nonlinear dynamic system were used to test. The results show that, compared with conventional nonlinear dynamic modeling methods, the proposed one possesses prominent advantages: 1) Only once dynamic calibrating experiment need be made; 2) The analytic expressions of nonlinear dynamic model are derived; 3) The model is more robust in noise resistance due to the good features of SVR. It provides a better way to model the nonlinear dynamic system.
��»�;Dehui Wu. ����SVR�ķ����Զ�̬ϵͳ��ģ�����о�[J]. �����Ӧ��, 2007, 27(9): 2253-2255.
Dehui WU. Research into modeling method of nonlinear dynamic system based on SVR. Journal of Computer Applications, 2007, 27(9): 2253-2255.