arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.04682cs.ITmath.IT
收藏

面向多用户跨近远场通信的宽间距阵列混合波束成形

Hybrid Beamforming with Widely-spaced-array for Multi-user Cross-Near-and-Far-Field Communications

Heyin Shen, Yuhang Chen, Chong Han, Jinhong Yuan

AI总结:

针对6G太赫兹密集多用户场景下角度相近用户难以区分的问题,本文采用MU-WSA构建跨近远场通信范式,优化阵列参数并提出AO与SVR混合波束成形算法,显著提升SE并降低功耗与计算时间。

摘要:

凭借数GHz带宽,太赫兹(THz)波束成形在第六代(6G)及未来通信中受到越来越多关注。现有波束成形设计主要聚焦于紧凑天线阵列,其典型通信发生在远场。然而,在密集多用户场景中,仅依赖远场角度域无法区分角度相近的用户。因此,本文采用多用户宽间距阵列(MU-WSA),通过扩大近场区域引入额外的距离域,从而形成跨近远场(CNFF)通信的新范式。在该范式下,研究了CNFF信道模型,并基于此优化MU-WSA中的子阵列间距$d_s$和子阵列数量$K$,以最大化信道容量。随后,在子连接系统中,针对模拟预编码器的特殊分块对角形式,提出了一种交替优化(AO)波束成形算法。对于全连接系统,通过构造MU-WSA的专用导向向量,提出了一种基于导向向量重构(SVR)的低复杂度算法。数值评估表明,得益于距离域分辨率,在20 dBm功率下,与紧凑阵列相比,MU-WSA可将频谱效率(SE)提升超过60%。此外,子连接(SC)系统中所提出的AO算法可实现全连接(FC)系统总SE的80%以上,同时将移相器数量减少$K^2$,从而降低功耗。全连接系统中的SVR算法可达到SE上界的95%以上,而运行时间仅为基于奇异向量分解(SVD)算法的10%。

英文摘要:

With multi-GHz bandwidth, Terahertz (THz) beamforming has drawn increasing attention in the sixth generation (6G) and beyond communications. Existing beamforming designs mainly focus on a compact antenna array where typical communication occurs in the far-field. However, in dense multi-user scenarios, only relying on far-field angle domain fails to distinguish users at similar angles. Therefore, a multi-user widely-spaced array (MU-WSA) is exploited in this paper, which enlarges the near-field region to introduce the additional distance domain, leading to a new paradigm of cross-near-and-far-field (CNFF) communication. Under this paradigm, the CNFF channel model is investigated, based on which the subarray spacing $d_s$ and the number of subarrays $K$ in MU-WSA are optimized to maximize the channel capacity. Then, in sub-connected systems, an alternating optimization (AO) beamforming algorithm is proposed to deal with the special block-diagonal format of the analog precoder. For fully-connected systems, a low-complexity steering-vector reconstruction (SVR)-based algorithm is proposed by constructing specialized steering vectors of MU-WSA. Numerical evaluations show that due to distance domain resolutions, the MU-WSA can improve the SE by over $60$% at a power of $20$dBm compared to the compact array. Additionally, the proposed AO algorithm in the SC system can achieve over 80% of the sum (SE) of the FC system while reducing the number of phase shifters by $K^2$, thereby lowering power consumption. The SVR algorithm in the FC system can achieve over 95% of the upper bound of SE but takes only 10% of the running time of the singular vector decomposition (SVD)-based algorithms.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑