Nearly optimal linear transceiver design for amplify-and-forward MIMO multiple-relay systems under MMSE criterion

Abstract

In this paper, we consider amplify-and-forward multiple-input multiple-output multiple-relay systems, where all the nodes have multiple antennas. For enhancing link reliability, we address the problem of designing optimal linear transceiver to minimize the mean squared error (MSE) of symbol estimations subject to the total relay transmit power constraint. This problem is highly complex and has not been solved in the literature. We first simplify this optimization problem to one that takes a singular value vector and a unitary matrix as optimization variables. Then based on the analyses for the simplified problem, we develop an iterative algorithm consisting of one boundary optimization and one unitary matrix constrained optimization. We show analytically that the proposed iterative algorithm always converges, and the MSE is monotonically decreasing from one iteration to the next. Finally, numerical results demonstrate the nearly optimal performance of the proposed scheme.