Blind equalization of polyphase FIR channels: a whitening approach

This article deals with the problem of blind estimation of linear equalizers based on the temporally or spatially oversampled channel output (SIMO problem). Algorithms are introduced and analyzed which elaborate on the whitening theorem, a simple result proposed in the context of blind SIMO equalization by Liu and Dong (see IEEE Trans. Circuits and Systems-I: Fundamental Theory and Applications, vol.44, no.5, 1997). Some advantages of the proposed approaches over existing related works are: (1) the design of the linear receiver does not require channel pre-estimation, (2) it is based on a simple decorrelation criterion that meets optimality in the noise free case, (3) it is robust to channel order mismatch, and finally, as usual in second-order methods, (4) it is not affected by the distance of the input distribution from Gaussianity (it works even with Gaussian or super-Gaussian sources). An adaptive algorithm is studied that is suitable for the processing of long data samples. A block-iterative counterpart is also introduced for the processing of short data sequences. The performances of these methods are evaluated over simulated data.