Linear Transceiver Design With Per-Antenna Power Constraints in Full-Duplex Multi-User MIMO Systems

We study a mean-squared-error-based linear transceiver design problem in a multiple-input multiple-output cellular system with a full-duplex base station serving multiple half-duplex uplink and downlink users simultaneously under the practical per-antenna power constraints. First, an iterative nonlinear Gauss-Seidel type algorithm updating the beamformer of each transmit antenna using simple closed-form expressions is proposed. Then, a lower complexity design is proposed based on alternating direction method of multipliers.