Fuzzy wavelet network modeling with B-spline wavelet
Jun Wang, Jian Xiao, Dan Hu · 2005
Based on the wavelet transform, we present a fuzzy wavelet network (FWN) for approximating arbitrary nonlinear function. Compared with the other methods, the FWN with B-spline wavelet has more frequency information. The approximation capability is proved. Simulation examples are given for static system and dynamic system respectively. The presented FWN not only reserves the multi-resolution capability of WNN, but also has the advantages of the approximation accuracy and good generalization performance.