MULTICHANNEL BLIND SOURCE SEPARATION AND BLIND EQUALIZATION USING FRACTIONAL SAMPLING

Yinglu Zhang and Saleem A. Kassam
E­mail: kassam@ee.upenn.edu

Blind source separation #BSS# in communications ap­ plications is usually considered under the assumption that the observed or received signals are instantaneous linear mixtures of the individual sources without in­ tersymbol interference #ISI#. Based on a more general structure for multi­input multi­output channels, we pre­ sent a method using fractional sampling and BSS to re­ cover transmitted symbols in the presence of ISI. This method is di#erent from traditional ways for dealing with ISI. It can be used successfully even if there are more sources than receivers. Simulation results are given, illustrating the good performance of the method and good agreementbetween theory and experiments. Keywords: Blind source separation, Intersymbol interference, Fractional sampling