Linear & non-linear channel equalization using Chebyshev artificial neural network

Abstract

In this paper, an efficient nonlinear artificial neural network (ANN) equalizer structure for channel equalization is proposed. The network structure utilizes expanding the input vector into the functional expansion block using expanded Chebyshev polynomials to form an enhanced high dimensional space. The structure of the Chebyshev neural network (ChNN) is used for equalization of linear and nonlinear channels. The Performance comparison of the ChNN has been carried out through extensive computer simulations and other neural networks equalizers. The result compared to show that ChNN provides superior performance in terms of convergence rate, computational complexity, MSE floor and BER over a wide range of channel conditions.