Performance of IA-MMSE Iterative Structures for SFBC Decoding in MIMO Systems Using Realistic System Parameters

Abstract

In this paper, a MIMO-OFDM system operating with SFBC encoded signals and affected by inter-cell interference is analyzed, and the performances of Interference Aware-Minimum Mean Square Error (IA-MMSE) symbol estimators are discussed in presence of realistic estimation of the interference auto-correlation matrix. Different estimation techniques are described and compared, based both on a system description using a $4\times 2$ channel matrix (requiring the estimation and inversion of a $4\times 4$ auto-correlation matrix), or using two $2\times 2$ matrices (requiring the estimation and inversion of two $2\times 2$ auto-correlation matrices). The first approach offers better performances, but the second approach has a very limited loss, counterbalanced by a simpler implementation complexity, characteristics that makes it an interesting solution in practical applications. An extensive simulation campaign using realistic system level parameters has been used for performance assessment.