Robust minimax equalization of imperfectly known frequency selective MIMO channels

Robust equalization of broadband channels with finite impulse responses (FIRs) and arbitrary number of inputs and outputs is considered. The robustness assumes addressing the problem of imperfect channel state information, present either at the receiver or at the transmitter. The erroneous channel knowledge is modeled by specified uncertainty regions. The goal is the optimization of the FER transceiver filtering with respect to various mean-square-error (MSE) based criteria, that must be satisfied for all disturbances that might appear, regarding the adopted error model. In particular, we analyze the robust counterparts of the problems of minimax MSE equalization, MSE fairness among spatial layers and MSE-constrained transmit precoding. Efficient numerical solutions for these problems are obtained using semidefinite programming methods from convex optimization theory.