Hybrid Beamforming Optimization for mmWave IRS-Aided MIMO systems

We study the joint optimization of a reflection vector and hybrid beamforming matrices for intelligent reflecting surface (IRS) assisted millimeter-wave (mmWave) multiple-input multiple-output (MIMO) systems with hybrid beamforming structures. Recently, an efficient and practical method to estimate the so-called effective (or cascaded) channel has been proposed. For the first time, we derive the near-optimal solution of the joint optimization only using the estimated effective channel, in which the modulus constraints of the reflection vector and analog beamforming matrices are satisfied asymptotically. By simply projecting our asymptotic one, we also derive the practical solution. Via simulations, it is demonstrated that our method can outperform the state-of-the-art (SOTA) method. Furthermore, the proposed method can ensure the robustness for inevitable channel estimation errors.