Adaptive self-organizing neural network method for tracking problems of nonlinear dynamic systems
Zhengzhi Wang, HU De-wen, Xiao Qi-Ying · 2002
In this paper a self-organizing neural network method of Kohonen (1990) and Matinez, Ritter, and Schulten (1989) is extended to solve the adaptive control problems of nonlinear dynamic systems. In each small region (receptive field), the nonlinear system can be expressed in linear vision approximately and controlled by neurons. The gaze points, receptive fields and control functions of neurons are regulated in self-organizing, learning and adaptive way. Several simulation examples verify the correctness and utility of this method.>