Abstract:
In this paper we design a self-organizing fuzzy neural network (SOFNN) by structure learning and parameter learning.A hybrid learning algorithm,by integrating back propagation and recursive least-squares (RLS) algorithm with forgetting factor,is used to learn the optimal parameters of the SOFNN.Furthermore,a fuzzy system is constructed and evaluated under the Schwarz & Rissanen information criterion (SRIC).Finally,several simulation examples of identification and model-reference tracking control of nonlinear systems are presented and analyzed to demonstrate the effectiveness of the proposed method,and the effect of the threshold parameter in fuzzy rule learning algorithm is also discussed.The simulation results show that this method can effectively prevent the system from overfitting,improve the generalization ability of the system,and acheive the control performance of the system.