智能反射面辅助的毫米波MIMO系统波束成形设计

智能反射面(Intelligent Reflecting Surface, IRS),也被称为可重构智能表面(Reconfigurable Intelligent Surface, RIS)或大型智能表面(Large Intelligent Surface, LIS),通过重新配置无线通信传输环境，可以有效提高通信质量。考虑IRS辅助的毫米波多输入多输出(Multiple-Input Multiple-Output, MIMO)下行系统，通过采用流形优化(Manifold Optimization, MO)技术处理，协同优化IRS的无源波束成形、功率分配、预编码矩阵以及合并矩阵，以最大化多用户和速率。由于利用流形优化算法设计的最佳目标函数是非凸的，因此利用毫米波级联信道的内在结构，将非凸问题转换为更加容易处理的问题。仿真结果表明，所采用的MO方案能获得比最小均方误差(Minimum Mean Square Error, MMSE)方案以及随机相位方案更好的和速率性能。

Beamforming design of millimeter wave MIMO system assisted by intelligent reflecting surface

Intelligent reflecting surface(IRS),also known as reconfigurable intelligent surface(RIS) or large intelligent surface(LIS),can effectively improve communication quality.In consideration of the IRS-assisted millimeter-wave multiple-input multiple-output(MIMO) downlink system,and through the use of manifold optimization(MO) technology processing,the passive beamforming,power allocation,precoding matrix and merging matrix of the IRS are collaboratively optimized to maximize multi-user sum-rate.Since the optimal objective function designed by the MO algorithm is non-convex,the internal structure of the millimeter-wave cascade channel is used to convert the non-convex problem into a problem easier to handle.The simulation results show that the adopted MO scheme can obtain better sum-rate performance than the minimum mean square error(MMSE) scheme and the random phase scheme.