智能反射面辅助的毫米波系统混合波束赋形设计

Hybrid Beamforming Design for Millimeter Wave System Assisted by Intelligent Reflecting Surface

将智能反射面(IRS)技术引入毫米波大规模多输入多输出(MIMO)系统,在保留各自优势的同时可以提高系统频谱效率.但现有研究主要是基于全数字波束赋形,每根天线需要连接一根专属的射频链,随着天线数增多会带来极大的硬件成本和功耗.针对这一问题,在IRS辅助的毫米波混合波束赋形MIMO系统,考虑功率约束和恒模约束,以频谱效率最大化为目标,提出了一种联合收发端和IRS波束赋形设计的低复杂度算法.首先通过数学变换得到系统频谱效率的近似表达式;然后,提出了一种迭代优化的方法对IRS相移矩阵进行设计;最后利用奇异值分解和流形优化技术优化混合预编码/合并矩阵.仿真了系统频谱效率随功率、IRS反射单元数及发射天线数变化的关系,结果表明,与交替方向乘子法方案和随机相位方案相比,本文方案在降低成本和功耗的同时,具有更高的频谱效率.

The introduction of intelligent reflecting surface(IRS)technology into millimeter-wave Massive Multiple-Input Multiple-Output(MIMO)systems can improve the spectral efficiency of the systems while retaining their respective advantages.However,the existing research is mainly based on full digital beamforming,and each antenna needs to be connected with a dedicated RF chain,with the increase of the number of antennas,it will bring great hardware costs and power consumption.To solve this problem,in the IRS-assisted millimeter wave hybrid beamforming MIMO system,considering the power constraints and constant modulus constraints,and aiming at maximizing spectral efficiency,a low-complexity algorithm for joint transceiver and IRS beamforming design is proposed.Firstly,the approximate expression of spectral efficiency of the system is obtained by mathematical transformation.Then,an iterative optimization method is proposed to design the IRS phase shift matrix.Finally,singular value decomposition and manifold optimization techniques are used to optimize the hybrid precoding/combining matrix.The relationship between the spectral efficiency of the system and the power,the number of IRS reflection elements and the number of transmit antennas is simulated.The results show that,compared with the alternative direction method of multipliers scheme and the random phase scheme,the proposed scheme has higher spectral efficiency while reducing cost and power consumption.