Abstract
We consider an amplify-and-forward multiple-input multiple-output relay system with a direct source-destination link. We adopt the minimum-mean-square- error (MMSE) criterion at the destination. The problem of interest is to jointly design source and relay precoders so as to minimize mean square error of transmitted symbols under total power constraints at the source and relay nodes. We propose a method which diagonalizes the MSE matrix using singular value decomposition (SVD) and generalized SVD techniques. The proposed approach based on this diagonalized MSE matrix is suboptimal and aims to reduce the design complexity of the precoders. The solution can be obtained via an iterative water-filling technique. Simulations results show the performance advantages of the proposed approach. © 2013 IEEE.