Fuzzy Neural Network Blind Equalization Algorithm Based on Radial Basis Function

Radial Basis Function (RBF) networks have simple network structures and fast convergence speed. In this paper, we propose a fuzzy neural network blind equalization algorithm based on radial basis function (FNN-RBF-BLE). The proposed algorithm defines the centers of RBF equalizer by analyzing the relationship of the input signal of equalizer and transmitted signal, therefore the structures of equalizer become simpler and the convergence speed become faster. Then, we use the fuzzy C-means clustering algorithm (FCM) divide the input signal of equalizer into each cluster center with different membership values and the mean square error (MSE) is reduced. The performance of the proposed algorithm is compared with blind equalization algorithm based on RBF (RBF-BLE). It is shown that a relatively low mean square error and fast convergence speed has been achieved.