Subspace-Based Methods For Blind Identification And Blind Equalization

In this paper, the problem of blind identification and blind equalization of digital communication systems based on the second-order statistics (SOS) are addressed. By analyzing, the corresponding methods are obtained.Firstly, this paper describes the subspace method of single-input multiple-output (SIMO) systems. This method can identify the transfer function of the system only using SOS, by sampling multiple outputs or oversampling a single output. It appears as a major breakthrough for allowing the blind identification and blind equalization only using SOS, and so the methods based on SOS become the important methods.Multiple-input multiple-output (MIMO) systems are mainly studied here, because MIMO systems are more likely to be applied. In this paper, the theories of rational subspace and minimal polynomial basis (MPB) are introduced. These theories can not only explain subspace method of SIMO systems but also offer an important support to the blind identification and blind equalization of MIMO systems.Finally, the subspace method of SIMO systems is extended to MIMO systems. A method is obtained by utilizing the theories of rational subspace. Note that this method is not simple extension because it needs to introduce the concept and techniques of blind separation.