Cluster-Based Multi-Carrier Hybrid Beamforming for Massive Device Terahertz Communications

Hang Yuan, Nan Yang, Xuhui Ding*, Chong Han, Kai Yang*, Jianping An

*Corresponding author for this work

    Research output: Contribution to journalArticlepeer-review

    22 Citations (Scopus)

    Abstract

    We propose a cluster-based multi-carrier beam division multiple access (MC-BDMA) scheme to enable massive device terahertz (THz) communications with the dynamic-subarray hybrid beamforming (HBF) architecture. This scheme is motivated by a limitation of conventional HBF architectures, i.e., the number of users cannot exceed the number of radio frequency (RF) chains. By exploiting the unique properties of THz channels, such as high distance-and-frequency dependence, high sparsity, and small angular spread, we propose multiple users in the same cluster to be supported by a single RF chain with distance-aware multi-carrier modulation. Moreover, we propose different clusters to be divided by BDMA that is accomplished through HBF design. In our proposed scheme, we first design a novel antenna selection and subarray partition algorithm for the dynamic-subarray HBF architecture to compensate for performance loss caused by diverse channels among users. We then design an algorithm for multi-carrier HBF which maximizes the achievable throughput and eliminates the inter-beam interference. Numerical results show that our proposed cluster-based MC-BDMA scheme with the dynamic-subarray HBF architecture provides an almost 100% spectral efficiency gain and a 60% energy efficiency gain over the conventional non-cluster HBF scheme in massive device communications.

    Original languageEnglish
    Pages (from-to)3407-3420
    Number of pages14
    JournalIEEE Transactions on Communications
    Volume70
    Issue number5
    DOIs
    Publication statusPublished - 1 May 2022

    Fingerprint

    Dive into the research topics of 'Cluster-Based Multi-Carrier Hybrid Beamforming for Massive Device Terahertz Communications'. Together they form a unique fingerprint.

    Cite this