On equalization with complex MLP network in the GSM environment

Neural networks have been studied for channel equalization purposes with quite promising results. However, not a lot of published results are available for their performance in realistic mobile systems, such as GSM (Global System for Mobile communications). Therefore, we have studied the use of a complex-valued multilayer perceptron (MLP) network, trained with a complex backpropagation (BP) algorithm, for equalization purposes in the GSM environment. Performance comparisons are made with a conventional decision-feedback equalizer (DFE) in terms of bit error rates and computational complexity.