Blind identification and equalization of multichannel FIR systems in unbalanced noise environments

This paper considers the problem of identifying and equalizing a set of FIR channels driven by the same input sequence and with outputs affected by unknown amounts of additive noise. The new procedure proposed here allows the estimation of the variances of the noises, of the order of the channels and of their parameters. This procedure leads to good results also when applied to short sequences of data as required in fast-varying environments like mobile communication ones.