抄録
High speed communication channels are usually impaired by nonlinear inter-symbol interference(ISI) and additive noise. In order to reconstruct the transmitted data sequence, the ISI must be compensated at the receiver, this is referred to as channel equalization. The blind channel equalization techniques should be done based only on the received sequence and known statistics of the input. In this paper, we propose a new blind method based on higher-order cumulants fitting for nonlinear channel, in which a 4th-order cumulant is used as a constraint instead of the 2nd-order statistic in order to smooth out the additive gaussian noise. Furthermore, in order to avoid convergence to local minima and obtain good convergence performance, a simplex-GA hybrid optimization is used to estimate the unknown parameter. Finally we use maximum likelihood sequence estimation(MLSE) equalizer for reconstructing input signals.
