Joint Optimization of the Worst-Case Robust MMSE MIMO Transceiver

We consider a robust MIMO transceiver design to minimize mean square errors, by taking into account the imperfect channel state information from a worst-case robustness perspective. We show that for a given precoder, the optimal robust equalizer is obtained through channel diagonalization, and vice versa. Thus, a joint but usually suboptimal transceiver design can be readily obtained through alternately optimizing the equalizer and precoder, where in each iteration only a scalar problem is to be solved. We further propose efficient algorithms for the scalar optimization problems.