Robust Transceiver Optimization in Downlink Multiuser MIMO Systems with Channel Uncertainty

The problem of transceiver optimization in multiuser multiple-input multiple-output downlink wireless systems is considered. The base station is assumed to possess only estimated, erroneous values of channel coefficients. The exact channels lie in uncertainty regions, specified by the Frobenius norms of the error matrices. An iterative optimization of the transmit and receive filters is performed with the goal of minimizing the total transmit power, while satisfying the users´ mean square error (MSE) targets for all channels from the uncertainty regions. Each iteration consists of two steps that can be equivalently rewritten as semidefinite programs with efficient numerical solutions. It is shown that the whole algorithm converges. The proposed framework can be applied for solving robust counterparts of several related MSE-optimization problems. The modifications of the proposed algorithms for accommodating box-like disturbances are analyzed, as well.