Deep Unfolding-Enabled Hybrid Beamforming Design for mmWave Massive MIMO Systems
Deep Unfolding-Enabled Hybrid Beamforming Design for mmWave Massive MIMO Systems
复制标题
DOI：
10.1109/icassp49357.2023.10096658
复制
发表时间：
2023-06
期刊：
影响因子：
--
通讯作者：
N. Nguyen;Mengyuan Ma;Nir Shlezinger;Y. Eldar;A. L. Swindlehurst;M. Juntti
中科院分区：
文献类型：
--
作者：
N. Nguyen;Mengyuan Ma;Nir Shlezinger;Y. Eldar;A. L. Swindlehurst;M. Juntti
混合波束形成(HBF)是毫米波(毫米波)通信系统的关键使能技术，但HBF的优化往往是非凸的和大尺寸的。在本文中，我们提出了一种基于深度展开的高效HBF方案，称为ManNet-HBF，它近似最大化了系统的频谱效率。该算法首先将最优数字波束形成器分解为模拟波束形成器和数字波束形成器，然后将矩阵分解问题转化为等价的极大似然问题，并利用轻量级深度神经网络MANNet对其模拟波束形成解进行矢量化和估计。数值结果表明，提出的MANNet-HBF方法具有接近最优的性能，与传统的基于模型的方法相当或更好，并且具有非常低的复杂度和快速的运行时间。例如，在有128个发射天线的仿真中，它达到了黎曼流形方案的98.62%的误码率，但速度快了13250倍。
Hybrid beamforming (HBF) is a key enabler for millimeter-wave (mmWave) communications systems, but HBF optimizations are often non-convex and of large dimension. In this paper, we propose an efficient deep unfolding-based HBF scheme, referred to as ManNet-HBF, that approximately maximizes the system spectral efficiency (SE). It first factorizes the optimal digital beamformer into analog and digital terms, and then reformulates the resultant matrix factorization problem as an equivalent maximum-likelihood problem, whose analog beamforming solution is vectorized and estimated efficiently with ManNet, a lightweight deep neural network. Numerical results verify that the proposed ManNet-HBF approach has near-optimal performance comparable to or better than conventional model-based counterparts, with very low complexity and a fast run time. For example, in a simulation with 128 transmit antennas, it attains 98.62% the SE of the Riemannian manifold scheme but 13250 times faster.