LAUSR

To enhance secure communications, we deploy the dual-polarized antenna arrays at communication nodes of the multi-input multioutput (MIMO) system, where the base station communicates with multiple legitimate users in the presence of an eavesdropper. We also adopt the dual-structured precoding in which a preprocessing matrix based on the polarized array spatial correlation and a linear precoding based on the instantaneous channel state information (CSI) are concatenated. We design this dual-structured multiuser linear precoding under three cases. In the first case, given perfect global CSI, the secrecy rate optimization problem is formulated and transformed into the weighted minimum mean square error (MSE) problem, which can be effectively solved by the block coordinate decent method. In the second case, where the eavesdropper's CSI is unavailable, an artificial noise is generated to confuse the eavesdropper by minimizing the information transmit power subject to a preset MSE threshold for the recovered confidential signals, which can be solved by an efficient iterative algorithm. In the third case of imperfect global CSI, the robust optimization for secure communications is performed by minimizing the largest received MSE among the users subject to the total transmit power constraint, which can be reformulated into a biconvex semidefinite programming problem and solved by an efficient alternating convex optimization. Simulation results are included to demonstrate the excellent performance of our proposed designs over the conventional single-polarized array-based designs, in terms of achievable secrecy rate, minimum transmit power, and the MSE of recovered confidential signals.