Multi-User Analog Beamforming in Millimeter Wave MIMO Systems Based on Path Angle Information
Multi-User Analog Beamforming in Millimeter Wave MIMO Systems Based on Path Angle Information
复制标题
DOI：
10.1109/twc.2018.2883279
复制
发表时间：
2019-01
影响因子：
10.4
通讯作者：
Lisi Jiang;H. Jafarkhani
中科院分区：
文献类型：
--
作者：
Lisi Jiang;H. Jafarkhani
我们的目标是设计一种用于下行多用户毫米波系统的仅模拟波束形成方案，以同时优化波束形成增益和用户间干扰。传统的模拟波束形成方案，例如波束选择方法，使用与信道的最强路径相对应的阵列响应向量来生成指向用户的波束。在多用户系统中，这样的方案将导致大的用户间干扰，特别是当用户位置接近时。在本文中，我们制定了一个多目标问题，以取得波束形成增益和用户间干扰之间的平衡。为了解决这个问题，我们首先使用加权和方法将多目标问题转化为单目标问题。然后，我们使用半定规划技术，使模拟波束形成与恒定幅度的约束，易于处理。此外，为了减轻信道估计和反馈量化误差的影响，我们设计了一个鲁棒的波束形成方案，以提供对不完美的信道信息的鲁棒性。首先，我们开发了一个信道误差模型的散射集群信道模型，它可以作为一个通用的毫米波信道的信道误差模型。然后，我们制定了一个多目标的问题，使用随机方法来抑制干扰，同时提高波束形成增益。仿真结果表明，我们提出的非鲁棒多用户模拟波束形成器优于传统的模拟波束形成方法时，信噪比高，我们提出的鲁棒波束形成器可以提供高达109%的改善，在和速率相比，波束选择方法。
We aim to design an analog-only beamforming scheme for downlink multi-user mm-wave systems to optimize the beamforming gain and the inter-user interference at the same time. Traditional analog beamforming schemes, such as the beam selection method, use the array response vector corresponding to the strongest path of the channel to generate a beam pointing to the user. In multi-user systems, such schemes will lead to large inter-user interference, especially when the users are closely located. In this paper, we formulate a multi-objective problem to strike a balance between the beamforming gain and the inter-user interference. To solve the problem, we first use the weighted-sum method to transform the multi-objective problem into a single-objective problem. Then, we use the semi-definite programing technique to make the analog beamforming with constant-magnitude constraints tractable. Furthermore, to alleviate the effects of the channel estimation and feedback quantization errors, we design a robust beamforming scheme to provide robustness against imperfect channel information. We first develop a channel error model for the scattering clustered channel model, which can serve as a general channel error model for the mm-wave channels. Then, we formulate a multi-objective problem using the stochastic approach to suppress the interference and enhance the beamforming gain at the same time. The simulation results show that our proposed non-robust multi-user analog beamformer outperforms the traditional analog beamforming method when the SNR is high and our proposed robust beamformer can provide up to 109% improvement in the sum-rate compared with the beam selection method.