APP下载

超大规模阵列的快速近场分层波束训练算法

2023-08-01钱海峰

现代信息科技 2023年11期

摘  要:文章研究了超大规模阵列(The Extremely Large-scale Array, XL-array)通信系统高效近场波束训练的设计。与远场波束只搜索最佳波束成形向量训练方法相比,近场波束训练需要在角域和距离域进行。为减少近场波束训练开销,提出一种新的两阶段近场分层波束训练方法。第一阶段采用传统远场码本和角域粗细波束相结合扫描的新方法来确定用户的候选角度。第二阶段使用近场极化域码本寻找用户的最佳有效距离。仿真结果表明,所提出的波束训练方法显著降低了近场穷搜的训练开销。

关键词:超大规模阵列;波束训练;波束成形;训练开销

中图分类号:TN820.15;TP391.9 文献标识码:A 文章编号:2096-4706(2023)11-0064-05

A Fast Near Field Layered Beam Training Algorithm for Extremely Large-scale Arrays

QIAN Haifeng

(School of Internet of Things Engineering, Jiangnan University, Wuxi  214122, China)

Abstract: In this paper, the design of efficient near field beam training for the extremely large-scale array communication system is studied. Compared to the far field beam only searching for the optimal beam forming vector training method, the near field beam training needs to be conducted in the angular and distance domains. In order to reduce the cost of near field beam training, a new two-stage near field layered beam training method is proposed. In the first stage, a new method combining traditional far-field codebooks and angle domain coarse and fine beam scanning is used to determine the candidate angles of users. In the second stage, the near-field polarization domain codebook is used to find the best effective distance for users. The simulation results show that the proposed beam training method significantly reduces the training cost of near field exhaustive search.

Keywords: extremely large-scale array; beam training; beam forming; training cost

0  引  言

近些年来,具有有源或无源面元的XL-array技术和太赫兹(Terahertz, THz)通信技术已成为第六代(Sixth-Generation, 6G)无线网络[1,2]中显著提高频谱效率的一种最有前途的技术。与第五代(Fifth-Generation, 5G)的大規模多输入多输出相比,6G的XL-array引入几个新的信道特性。具体而言,随着天线数量和载波频率的显著增加,如果用户距离小于距离增大后的瑞利距离,用户更有可能定位在球形波前传播的近场区域。因此,基于远场信道模型和平面波前假设的传统收发器设计方法可能不再适用,需要专门设计适用于近场通信[3-5]的新方法。

在近场通信中,直接将传统的远场波束训练方法[6,7]用于近场通信系统将会导致性能大幅度下降,特别是当用户非常接近超大规模阵列的时候。因此,文献[5]中提出一种新的用于近场的极域码本,即对角域进行均匀采样,对距离域进行非均匀采样,以减小码本的尺寸。……

登录APP查看全文