Source-linked AI summary

Hybrid Beamforming for Terahertz Wireless Communications: Challenges, Architectures, and Open Problems

Chong Han, Longfei Yan, Jinhong Yuan

arXiv:2101.08469v1cs.ITeess.SPmath.IT

TL;DR

THz wireless offers broad bandwidth for 6G but faces sparse channels, blockage, large-array power demands, and propagation and beam-squint challenges. This paper analyzes these issues and evaluates three THz-specific hybrid-beamforming architectures; its reported evaluations include a 5.49 dB beam-squint array-gain loss.

  • Problem

    THz hybrid beamforming must address sparse channels, blockage, large-array constraints, beam squint, and spherical-wave propagation while balancing data rate and power consumption.

    Source: B. Limitations and Remaining Problems · p. 3, §VII · p. 7
  • Method

    The paper analyzes THz-specific challenges and evaluates WSMS, DAoSA, and TTD-based hybrid-beamforming architectures, alongside open problems and research directions.

    Source: §VII · p. 7
  • Results

    The evaluations report WSMS data-rate gains over FC and AoSA at 20 dBm, adjustable DAoSA data-rate and power, and a beam-squint array-gain loss as large as 5.49 dB.

    Source: §V · pp. 5–6
  • Takeaways & Limitations

    The paper concludes that THz hybrid-beamforming design has multiple open problems and potential research directions.

    Source: §VII · p. 7

    FC and AoSA do not balance power consumption and data rate in the THz band, and their spatial multiplexing gain remains bounded by low SDoF.

    Source: B. Limitations and Remaining Problems · p. 3

Abstract

from arXiv · show

Terahertz (THz) communications are regarded as a pillar technology for the sixth generation (6G) wireless systems, by offering multi-ten-GHz bandwidth. To overcome the short transmission distance and huge propagation loss, ultra-massive (UM) MIMO systems that employ sub-millimeter wavelength antennas array are proposed to enable an enticingly high array gain. In the UM-MIMO systems, hybrid beamforming stands out for its great potential in promisingly high data rate and reduced power consumption. In this paper, challenges and features of the THz hybrid beamforming design are investigated, in light of the distinctive THz peculiarities. Specifically, we demonstrate that the spatial degree-of-freedom (SDoF) is less than 5, which is caused by the extreme sparsity of the THz channel. The blockage problem caused by the huge reflection and scattering losses, as high as 15 dB or over, is studied. Moreover, we analyze the challenges led by the array containing 1024 or more antennas, including the requirement for intelligent subarray architecture, strict energy efficiency, and propagation characterization based on spherical-wave propagation mechanisms. Owning up to hundreds of GHz bandwidth, beam squint effect could cause over 5~dB array gain loss, when the fractional bandwidth exceeds 10%. Inspired by these facts, three novel THz-specific hybrid beamforming architectures are presented, including widely-spaced multi-subarray, dynamic array-of-subarrays, and true-time-delay-based architectures. We also demonstrate the potential data rate, power consumption, and array gain capabilities for THz communications. As a roadmap of THz hybrid beamforming design, multiple open problems and potential research directions are elaborated.

I. INTRODUCTION

THz offers the bandwidth envisioned for 6G data rates, but propagation loss and the scale of UM-MIMO arrays create design constraints. The paper examines these challenges and hybrid beamforming’s role, then surveys THz-specific architectures and open problems.

I. INTRODUCTION

  • THz offers several tens of GHz of consecutive bandwidth for 6G, while mmWave’s limited consecutive bandwidth makes supporting a 1 Tbps peak data rate difficult.The paper identifies the THz band as spanning 0.1–10 THz.Source: §I · p. 1
  • Huge propagation loss limits THz communication distance, while ultra-massive arrays such as 1,024 antennas can generate narrow beams and high array gain.The loss is attributed to the small effective area of THz antennas.Source: §I · p. 1
  • Hybrid beamforming combines limited RF chains with reduced hardware complexity and power consumption, while aiming for data rates comparable to fully digital beamforming.Fully digital designs require a dedicated RF chain and DAC/ADC per antenna; analog designs support only one data stream.Source: §I · p. 1
  • The paper examines low SDoF, blockage, large-array demands, spherical-wave propagation, and beam squint, then presents three THz-specific architectures and research directions.It reports that beam squint can cause over 5 dB of array-gain loss when fractional bandwidth exceeds 10%.Source: §I · pp. 1–2

II. CHALLENGES OF THZ HYBRID BEAMFORMING

THz channel sparsity restricts spatial multiplexing, and strong line-of-sight dominance makes links vulnerable to blockage. These channel properties constrain achievable data rate, including when UM-MIMO provides high array gain.

1) Channel sparsity and low SDoF:

  • Typically fewer than 5 multipath components bound SDoF and spatial multiplexing, so high array gain alone does not remove the resulting data-rate constraint.The passage attributes the sparse channel to THz scattering and diffraction losses.Source: 1) Channel sparsity and low SDoF · p. 2

2) Blockage problem:

  • Reflection paths can be over 15 dB weaker than the LoS path, making a blocked LoS link vulnerable to insufficient data rate or outage.The paper describes this blockage problem as more severe than at mmWave frequencies.Source: 2) Blockage problem · p. 2

B. Challenges from THz UM-MIMO Systems Perspectives

THz UM-MIMO arrays face power and modeling demands, while wide fractional bandwidth makes phase-shifter beamforming vulnerable to beam squint. Near-field links also require spherical-wave propagation to be considered.

1) Large-scale antenna array:

  • More than 1,024 antennas and high per-device power consumption can produce system consumption of several tens of watts, making power reduction important alongside high data rates.The cited examples give 60 mW for a power amplifier and 42 mW for a phase shifter.Source: 1) Large-scale antenna array · p. 2

2) Beam squint effect:

  • At THz fractional bandwidths, frequency-independent phase shifters cause phase errors away from the tuned frequency and can reduce array gain by about 5 dB.Fractional bandwidth is bandwidth divided by central frequency; the passage gives 50 GHz/300 GHz = 16.7% as an example.Source: 2) Beam squint effect · p. 2

3) Spherical-wave propagation:

  • At 1 THz, the Rayleigh distance for a 0.1 m array is 67 m, so THz links may fall within the region where planar-wave assumptions are invalid.The passage says spherical-wave propagation must then be considered, complicating channel modeling and estimation and influencing hybrid beamforming.Source: 3) Spherical-wave propagation · p. 2

III. TRADITIONAL HYBRID BEAMFORMING ARCHITECTURES AND LIMITATIONS

Traditional hybrid architectures trade hardware cost and power against data-rate capability. The literature addresses beamformer design through optimization and approximation methods, but these architectures have distinct implementation costs.

A. Existing Hybrid Beamforming Architectures and Algorithms

  • Fully connected (FC) links each RF chain to all antennas, requiring a number of phase shifters equal to antennas times RF chains, whereas AoSA uses one per antenna.AoSA connects each RF chain to a subset of antennas, reducing hardware complexity and power relative to FC.Source: A. Existing Hybrid Beamforming… · p. 3
  • Design algorithms maximize data rate directly or minimize distance from a fully digital beamformer, using methods including two-step optimization, SIC, OMP, and alternating optimization.The passage describes these approaches for FC and AoSA architectures.Source: A. Existing Hybrid Beamforming… · p. 3

B. Limitations and Remaining Problems

The paper identifies shortcomings in fully connected (FC) and array-of-subarrays (AoSA) designs and presents THz-specific architectures aimed at distinct design challenges. WSMS uses widely spaced subarrays to increase multiplexing, while its subarray count must balance multiplexing and array gain.

B. Limitations and Remaining Problems

  • FC offers higher data rate but high power consumption and hardware complexity, whereas AoSA reduces these costs at the expense of data rate; neither resolves low SDoF or beam squint.The paper therefore calls for THz-specific architectures that balance power and rate while addressing these unresolved constraints.Source: B. Limitations and Remaining Problems · p. 3

IV. THZ-SPECIFIC HYBRID BEAMFORMING ARCHITECTURES

  • The proposed WSMS, DAoSA, and TTD-based architectures target low SDoF, power–data-rate balance, and beam squint, respectively.Source: §IV · p. 3

A. THz WSMS Hybrid Beamforming Architecture

  • Widely spaced subarrays enable intra-path multiplexing alongside inter-path multiplexing, improving total multiplexing gain by a factor of the number of subarrays over FC and AoSA.The approach uses critically spaced antennas within each subarray and widely spaced subarrays; the resulting spherical-wave subpaths have distinct phases.Source: A. THz WSMS Hybrid Beamforming… · pp. 3–4
  • Increasing WSMS multiplexing gain reduces array gain and can affect SNR and data rate, so the number of widely spaced subarrays requires careful design.Source: A. THz WSMS Hybrid Beamforming… · p. 4

B. THz DAoSA Hybrid Beamforming Architecture

DAoSA uses switches to adjust connections between RF chains and subarrays, allowing different power-consumption and data-rate levels. Its design involves optimizing both the switch network and beamforming matrices, with algorithmic and hardware complexity remaining concerns.

B. THz DAoSA Hybrid Beamforming Architecture

  • Switches dynamically control RF-chain-to-subarray connections, with FC and AoSA as DAoSA’s all-switches-closed and one-connection-per-RF-chain special cases.Phase shifters behind open switches are inactive and do not consume power, allowing the design to balance power consumption and data rate.Source: B. THz DAoSA Hybrid Beamforming… · p. 4
  • DAoSA design alternates switch-network selection and beamforming-matrix design, using a tractable sequence for the switch problem and an element-by-element algorithm for irregular matrix structure.Source: B. THz DAoSA Hybrid Beamforming… · pp. 4–5
  • DAoSA still faces high algorithmic complexity from jointly designing switches and beamforming, as well as additional hardware burden from its switches.The paper identifies lower-complexity switch selection and a unified joint-design framework as potential directions.Source: B. THz DAoSA Hybrid Beamforming… · p. 5

C. THz TTD-based Hybrid Beamforming

TTD-based hybrid beamforming addresses wideband beam squint by replacing frequency-independent phase shifters with frequency-dependent delays. The approach introduces codebook-complexity and hardware-resolution trade-offs.

C. THz TTD-based Hybrid Beamforming

  • TTD phase adjustment can mitigate beam squint across FC, AoSA, WSMS, and DAoSA architectures.Its frequency-dependent phase shift can match the required wideband phase shift.Source: C. THz TTD-based Hybrid Beamforming · p. 5
  • The codebook-based design in redesigns the phase-shifter codebook for TTD frequency dependence, but large codebooks raise complexity.Source: C. THz TTD-based Hybrid Beamforming · p. 5
  • Infinite-resolution TTD can eliminate beam squint without array-gain loss, but ideal or high-resolution devices are power-hungry and complex.Low-resolution TTD is more suitable for practical systems, but its limited resolution can leave residual beam squint; the resolution–gain-loss relationship and low-complexity non-codebook algorithms remain open issues.Source: C. THz TTD-based Hybrid Beamforming · p. 5

V. PERFORMANCE EVALUATION

The evaluation compares WSMS, DAoSA, and TTD-based beamforming through data-rate, power-consumption, and array-gain simulations. The results illustrate WSMS data-rate gains, DAoSA rate–power adjustment, and TTD's resistance to beam-squint loss.

V. PERFORMANCE EVALUATION

  • 40 Gbps and 60 Gbps higher data rates than FC and AoSA, respectively, were reported for WSMS at 20 dBm.With two widely spaced subarrays, WSMS doubles SDoF and multiplexing gain, addressing the low-SDoF problem.Source: §V · p. 5
  • With 4–16 closed switches, DAoSA adjusts data rate and power consumption between AoSA-like minima and FC-like maxima.The simulation uses 4 RF chains, 16 switches, 5 GHz bandwidth, and 20 dBm transmit power.Source: §V · p. 5, Figure 4 · p. 6
  • 5.49 dB array-gain loss occurs with phase-shifter beam squint as frequency departs from the center; TTD maintains constant array gain across carrier frequencies.The comparison uses 30 GHz bandwidth and 30 subcarriers.Source: §V · pp. 5–6, Figure 5 · p. 6

A. Hardware-efficient Design

The open problems span hardware efficiency, imperfect channel knowledge, and computational complexity. Proposed directions include low-resolution hardware-aware designs, improved channel estimation and robust algorithms, and deep-learning methods adapted to THz challenges.

A. Hardware-efficient Design

  • THz hardware design must account for difficult-to-produce high-resolution phase shifters, TTDs, and DAC/ADCs; lens arrays avoid phase shifters or TTDs but require extensive beam-selection switches.Hardware-efficient lens-array architectures and algorithms are still needed.Source: A. Hardware-efficient Design · p. 6

B. Influence of Imperfect Channel State Information

  • High-dimensional channels make perfect CSI difficult to obtain, while sharp beams make imperfect CSI and misalignment especially damaging.Suggested directions include super-resolution channel and angle estimation and algorithms robust to imperfect CSI.Source: B. Influence of Imperfect Channel… · p. 6

C. Deep Learning Algorithms for Hybrid Beamforming

  • Existing hybrid-beamforming algorithms can scale with the number of antennas or its square, making computation costly for ultra-massive arrays.Deep-learning methods are a potential direction, but THz-specific features and challenges must be considered.Source: C. Deep Learning Algorithms for Hybrid… · p. 6

D. Joint Active and Passive Beamforming

Joint active and passive beamforming is proposed as a direction for mitigating THz blockage, using transmitter/receiver arrays with an intelligent reflecting surface. The conclusion situates this direction within a broader review of THz-specific challenges, architectures, simulations, and open problems.

D. Joint Active and Passive Beamforming

  • When LoS is blocked, weak remaining paths can cause outage; an IRS could strengthen reflected signals alongside active beamforming.Performance analysis and joint active–passive beamforming design for the THz band remain open research needs.Source: D. Joint Active and Passive Beamforming · pp. 6–7

VII. CONCLUSION

  • The paper's scope covers low SDoF, blockage, large arrays, beam squint, and spherical-wave propagation, alongside three THz-specific architectures and open problems.Simulations address data rate, power consumption, and array gain.Source: §VII · p. 7
Loading 2101.08469v1…