Neural network techniques applied to channel equalization in indoor radio environment

Indoor communications are affected by multipath propagation that imposes an upper limit on the system data rate and restricts their possible applications. In order to overcome the degrading effect introduced by the channel, conventional equalizers, based in digital filters, have been traditionally used. In this paper a new approach based on neural networks is considered. In particular the behavior of multilayer perceptron (MLP) structures applied to the channel equalization of cellular radio indoor environments for 4-QAM modulation is analyzed. In this case the system performance becomes limited by the co-channel interference rather than by the noise and, again, the equalization techniques based on MLP structures may be an effective tool in order to fight this degradation effect