Abstract:

A fuzzy neural network (FNN) based fault diagnosis model for the code division multiple access (CDMA) network is proposed which focusing on solving the difficult problem that general diagnosis algorithms can hardly build model for the CDMA network fault diagnosis system. For the proposed fault diagnosis model, the numbers of neurons in the output layer are based on the fault types of the CDMA network, and the amount of input variables needed for those fault diagnosis determine the number of neurons in the input layer. Moreover, the synaptic weighting value and fuzzy member function are obtained from training the model with the input/output fault diagnosis data in the expert knowledge base. Computer simulation results show the effectiveness and the applicability of the proposed method.

Key words: fault diagnosis, fuzzy-neural network, code division multiple access network, network optimization