Channel equalization for severe intersymbol interference and nonlinearity with a radial basis function neural network

RBF neural network has been applied to equalization for a channel with severe intersymbol interference and nonlinearity expressed with a 3rd order Volterra series. Performances of three different equalizers are compared for this channel. They are: a linear transversal equalizer, a feedforward network equalizer with sigmoid neurons, and a radial basis function network equalizer. According to the simulation results, the radial basis function neural network equalizer achieved a much faster convergence rate and superior performance than the other equalizers.