Blind Identification of FIR Systems and Deconvolution of White Input Sequences

This paper deals with the problem of blind identification and deconvolution of FIR channels driven by a white input sequence with unknown variance, in an unbalanced noise environment. By using the structural properties of the covariance matrix of the input, an estimate of the channel coefficients is obtained. The subsequent deconvolution of the unknown input signal is then performed by means of a minimum variance equalizer. The performance of the proposed algorithm is tested by means of Monte Carlo simulations. Key-Words: Blind identification, blind equalization, multichannel FIR systems, white input sequences.