RADIAL BASIS FUNCTION EQUALIZATION USING COMPETITIVE LEARNING

learning In this paper the problem of equalization of multiple quadrature amplitude modulated signals, using a radial basis function (RBF) neural network, is studied. Because the equalizer performance is directly related to the estimations of the RBF centres, different competitive learning algorithms for the RBF centres are presented. A new competitive algorithm is introduced, the rival penalized competitive learning, which rewards the winner and penalises its first rival. Simulations results, performed in different conditions, are presented, showing that the performance of the RBF equalizer, if trained with this new algorithm is better comparative with other competitive algorithms. 1