毫米波大规模MIMO系统中的一种低复杂度的混合波束赋形方案

A Low Complexity Hybrid Beamforming Scheme for Millimeter-Wave Massive MIMO System

针对传统模拟/数字混合波束赋形需要进行信道估计和高维奇异值分解(Singular Value Decomposition,SVD)的问题,设计了一种低复杂度的混合波束赋形方案。在模拟域,采用离散傅立叶变换(Discrete Fourier Transform,DFT)设计出码本波束,并提出一种基于期望信号的波束扫描方法,搜索出最接近期望用户方向的模拟波束。在数字域,结合自适应算法将波束微调对准期望用户方向,并在干扰方向形成较深的“零陷”。通过理论分析和仿真,对误差收敛、信干噪比以及频谱效率等性能进行了研究。结果表明,所设计的混合波束赋形方案在不涉及信道估计、SVD分解的情况下,信干噪比和频谱效率等性能仍可以接近于纯数字波束赋形方案。

The traditional analog/digital hybrid beamforming technology requires channel estimation and high-dimension SVD(Singular Value Decomposition)operations.To avoid these complex operations,a lowcomplexity hybrid beamforming scheme combining beam scanning and adaptive algorithm is designed.Firstly in analog domain,DFT(Discrete Fourier Transform)is used to design codebook beams,and then a beam scanning method based on the desired signal is proposed so as to find the analog beam closest to the desired direction.Then in digital domain,combined with adaptive algorithm,the beam is fine-tuned to the desired user direction and a deeper“null”is formed in the interference direction.Through theoretical analysis and simulation,the performance of EC(Error Convergence),SINR(Signal to Interference plus Noise Ratio)and SE(Spectral Efficiency)is explored.Simulation results show that the designed hybrid beamforming scheme does not involve channel estimation and SVD decomposition,and its performance such as SINR and SE can still be close to the pure digital beamforming scheme.