Separate MMSE Optimization of Multi-Hop Non-Regenerative MIMO Relay Communication

This paper addresses to a problem of the minimum mean-squared error (MMSE) optimization of transmitters and receivers in a multi-hop multiple-input multiple-output (MIMO) non-regenerative relay system. Instead of global search methods, we propose a separate strategy for linear precoding and decoding matrices design, which divides the total optimization problem into a set of consecutively solved subproblems. For each subproblem a closed-form water-filling type solution can be easily obtained. After optimization the overall link between a source and a destination is diagonalized. Our approach reduces the number of auxiliary links to feedback channel state information. The simulation results demonstrate high performances of the proposed strategy.