NOVEL APPROACH TO COCHANNEL INTERFERENCE MINIMIZATION USING WILCOXON MULTILAYER PERCEPTRON NEURAL NETWORK

Abstract

This paper is based on, a novel wilcoxon learning machines technique used in artificial neural network equalizer, to mitigate the co-channel interference in the presence of additive white Gaussian noise (AWGN). Multilayer Perceptron Network, Radial Basis Function Network,Reurrent network, Functional link Artificial Neural Networks, Chebyshev Neural Networks, Fuzzy and Adaptive Neuro Fuzzy System have been successfully used in equalization. In this paper we proposed a MLP based equalizer trained using wilcoxon learning named as wilcoxon Multilayer Perceptron Neural network (WMLPNN). WMLPNN is a rank based statistics approach, in which weights and parameters of the network are updated using rules based on gradient descent principle. The Performance of the WMLPNN has been demonstrated through extensive computer simulations and compared with other neural networks equalizers in terms of convergence rate, computational complexity and bit error rate. The experiments results show the proposed method can effectively solve this class of problem.