BLIND FRACTIONALLY-SPACED EQUALIZATION, PERFECT-RECONSTRUCTION FILTER BANKS AND MULTICHANNEL LINEAR PREDICTION

Abstract

Equalization for digitd conimunicatiou~ constitutes a very particular blind deconvoliition problem in that the received signal is cyclostationary. Oversampling (OS) (w.r.t. the symbol rate) of the cyclostationary received signal leads to a stationary vect.or-valued signal (polyphase representation (PR)). OS also leads to a fractionally-spa.ced channel model and equalizer. In the F'R, channel and equalizer can be considered as an analysis and synthesis filter hank. Zero-forcing p equalization corresponds to a perfect-reconstruction Iter bank. We show that in the OS (:aye FIR ZF equalizers exist for a FIR channel. In the PR, the multichannel linear prediction of the noiseless received signal becomes singular eventually, reminiscenc of the singlechannel prediction of a sum of sinusoids. A:. a result, the chaiinel can be identified from the received signal second-order statistics by linear prediction in the noise-free case, and by rising the Pisarenko method when there is a,dditive noise. In the given data case, Music (subspace) or MI. te(hniqiies can be applied.