Learning Algorithm of Parameters about Fuzzy Membership Functions Based on the RBF Neural Network

摘要

In order to solve the current development difficulty of fuzzy control system - how to find the most optimal membership functions, in this paper, an improved RBF neural network structure used to extract the fuzzy rules and a learning algorithm of the parameters of fuzzy Membership functions based on this network are discussed. Make full use of the learning ability of the RBF neural network, membership functions are found from the historical data. To some extent, this solution decreases the correspondent work of system's development and overcomes some errors probably caused by lack of experience. Finally, using VC++ and MATLAB language, the simulation experiment proves that this algorithm is effective.