Semiblind multiuser MIMO channel estimation using capon and MUSIC techniques

Channel state information (CSI) is required at the receiver side of multiple-input multiple-output (MIMO) systems for decoding space-time codes. In this paper, new semiblind techniques are developed to jointly estimate the receiver CSI of multiple multiantenna transmitters that use orthogonal space-time block codes (OSTBCs). The proposed techniques are based on the extension of the concepts of the popular Capon and MUSIC spectral estimators to the problem of multiuser channel estimation. Our methods blindly estimate the subspace that contains the user channel matrices, and then use only a few training symbol blocks to extract the user CSI from this subspace. It is demonstrated that, in addition to improving the bandwidth efficiency, the proposed techniques offer better channel estimation accuracy as compared with the standard (nonblind) least-squares (LS) channel estimation method. Moreover, it is shown that using the proposed Capon- and MUSIC-based semiblind channel estimators along with the coherent maximum-likelihood (ML) or the minimum-variance (MV) multiuser MIMO receivers results in a substantially better performance as compared with the case when the standard LS channel estimates are used