Simplified alternating optimization for robust MMSE MIMO transceiver

Abstract

We consider a worst-case robust transceiver design for multiple-input multiple-output (MIMO) channels, under the minimum mean square error (MMSE) criterion, to combat against the imperfection of channel state information (CSI). The joint robust transceiver design is formulated as a minimax problem. We show that for a given precoder (equalizer), the optimal robust equalizer (precoder) is obtained in a semi-closed form by channel diagonalization. Thus, a joint robust transceiver can be efficiently achieved by alternately optimizing the equalizer and precoder, where in each iteration only a power allocation problem is to be solved. We also propose efficient algorithms to solve the power allocation problems in each iteration.