Robust MSE-constrained downlink precoding in multiuser MIMO systems with channel uncertainty

We study robust transceiver optimization methods for a multiuser downlink wireless system. The robustness assumes addressing the problem of the imperfect channel state information (CSI) at the base station (BS). The imperfect knowledge is described by the specified uncertainty regions, known to the transmitter, where the accurate channel values are situated. The BS and the users are equipped with antenna arrays. The complete transmit filter at the BS and the scaling factors of the fixed receive filters are jointly optimized to minimize the total transmit power under the set of users' mean square error (MSE) targets. The MSE constraints are supposed to be obeyed for all channels that belong to the corresponding uncertainty regions. The convexity of the problem in general is shown and, for two characteristic cases of the CSI disturbances, efficient numerical solutions, based on semidefinite programming methods, are derived. For the fixed part of the receive filters, some heuristic solutions are proposed. Various further applications of the proposed framework are shortly discussed, as well.