Source-linked AI summary

Ultra-Massive MIMO Systems at Terahertz Bands: Prospects and Challenges

Alice Faisal, Hadi Sarieddeen, Hayssam Dahrouj, Tareq Y. Al-Naffouri, Mohamed-Slim Alouini

arXiv:1902.11090v4eess.SP

TL;DR

THz communications offer large bandwidths for future 6G data-rate demands, but propagation losses and power limitations constrain communication range. The paper surveys THz UM-MIMO transceiver and channel considerations, research advances, and prospective use cases. It argues that properly configured UM-MIMO arrays can overcome distance and power limitations.

  • Problem

    THz communications promise large bandwidths for future 6G data rates, but high propagation losses and power limitations constrain communication range.

    Source: §I · pp. 1–2
  • Method

    The paper links THz UM-MIMO transceiver design, channel characteristics, research advances, and prospective use cases in a holistic overview.

    Source: §I · p. 2
  • Results

    The paper argues that properly configured UM-MIMO arrays can overcome distance and power limitations, and that graphene-based nano-antenna arrays can efficiently realize THz UM-MIMO.

    Source: §VI · p. 7
  • Takeaways & Limitations

    The paper identifies prospective THz UM-MIMO uses in communications and sensing, data centers, reconfigurable intelligent surfaces, and mobile mid-range communications.

    Source: §V · p. 6

Abstract

from arXiv · show

Terahertz (THz)-band communications are currently being celebrated as a key technology that could fulfill the increasing demands for wireless data traffic in the upcoming sixth-generation (6G) of wireless communications. Many challenges, such as high propagation losses and power limitations, which result in short communication distances, have yet to be addressed for this technology to be realized. Ultra-massive multiple-input, multiple-output (UM-MIMO) antenna systems have emerged as practical means for combatting this distance problem, thereby increasing system capacity. Towards that end, graphene-based nano-antennas have recently been proposed, as they can be individually tuned and collectively controlled in compact UM-MIMO array-of-sub-arrays architectures. In this paper, we present a holistic overview of THz UM-MIMO systems. We assess recent advancements in transceiver design and channel modeling, and discuss the major challenges and shortcomings of such designs by deriving the relationship between communication range, array dimensions, and system performance. We further highlight several research advances that could enhance resource allocation at the THz band, including waveform designs, multi-carrier configurations, and spatial modulations. Based on this discussion, we highlight prospective use cases that can bring THz UM-MIMO into reality in the context of sensing, data centers, cell-free systems, and mid-range wireless communications.

I. INTRODUCTION

THz bandwidth may meet future 6G data-rate demand, [3], but high propagation losses and power limitations constrain coverage. This paper links THz UM-MIMO transceiver design, channel conditions, and prospective use cases.

I. INTRODUCTION

  • THz communications are proposed to exploit large bandwidths for the growing data-rate demands of future 6G wireless communications and beyond, [3].The passage contrasts this with mmWave systems explored over 30 to 300 GHz.Source: §I · p. 1
  • High-frequency THz devices face implementation hurdles: electronic generators and detectors perform poorly at room temperature, while photonic solutions have limited output power.The paper proposes integrated electronic and photonic solutions that are programmable and efficient, though not necessarily perfect.Source: §I · pp. 1–2
  • Graphene’s electrical tunability and configurability support high-frequency signals, while its plasmonic antennas enable compact arrays with many antenna elements,.Their surface plasmon polariton wavelength is much smaller than the electromagnetic wavelength.Source: §I · p. 2
  • UM-MIMO can increase communication range and improve achievable data rate through spatial multiplexing and beamforming, respectively [4].The paper presents UM-MIMO as one of two alternative means, alongside reflect arrays, to extend THz coverage.Source: §I · p. 2
  • The paper aims to connect THz UM-MIMO transceiver design and channel characteristics with prospective use cases, addressing a gap in holistic coverage of these topics.It outlines array design and channel modeling, then identifies research advances and prospective use cases.Source: §I · p. 2

II. ARRAY OF SUB-ARRAYS DESIGN

The section explains how array-of-sub-arrays (AoSA) designs trade compactness against beamforming and multiplexing gains, and compares multiplexing and power considerations for mmWave and THz arrays.

II. ARRAY OF SUB-ARRAYS DESIGN

  • AoSA partitions a large array into sub-arrays: elements within each sub-array improve beamforming, while collaboration among sub-arrays provides spatial-multiplexing gain.This arrangement can provide directed independent paths that carry independent information.Source: §II · p. 2
  • For same-size arrays and matched transmit power, mmWave and THz yield identical multiplexed streams, although achievable gains depend on channel conditions.The comparison assumes a single dominant LoS ray; spherical-wave propagation at THz could improve gains.Source: §II · p. 2, Figure 2 · p. 3
  • True THz-component power metrics are lacking, so the power analysis estimates consumption from state-of-the-art component characteristics in a single-carrier system.The considered components are the LNA per antenna element, ADC per sub-array, and mixer per sub-array.Source: §II · p. 3
  • 140 GHz receiver power is much higher than 28 GHz receiver power for both 1 cm² and 10 cm² footprints, according to the analysis.The estimate considers LNA power per antenna element and ADC and mixer power per sub-array.Source: §II · p. 3, Figure 3 · p. 3

III. CHANNEL MODELING AND CHARACTERISTICS

THz UM-MIMO performance depends on channel conditions and channel-state information, making accurate channel models a prerequisite for efficient use of the band.

III. CHANNEL MODELING AND CHARACTERISTICS

  • The section identifies accurate channel models as a prerequisite for efficient THz use because system performance depends on channel conditions and channel-state information.Source: §III · p. 4
  • Models should account for spreading and molecular-absorption losses and include LoS, NLoS, reflected, and scattered paths.Source: §III · p. 4
  • The modeling discussion distinguishes static from time-varying environments and reviews approaches alongside peculiar THz-channel characteristics.Source: §III · p. 4

A. THz Channel Modeling

THz channel models may be deterministic, statistical, or hybrid, and large arrays introduce non-stationarity and challenges to the plane-wave assumption.

A. THz Channel Modeling

  • THz channel modeling approaches are deterministic, statistical, or hybrid.Hybrid models capture dominant paths deterministically and generate other paths statistically.Source: A. THz Channel Modeling · p. 4
  • Deterministic models use site geometry and ray tracing, while statistical models use matrix-based or reference-antenna-based propagation representations.Point-to-point ray tracing can be mapped to other antenna elements to reduce computational complexity.Source: A. THz Channel Modeling · p. 4
  • Large arrays can exhibit non-stationarity across regions and make the plane-wave assumption difficult to satisfy, motivating spherical-wave approaches.The passage particularly notes this plane-wave challenge for non-plasmonic arrays.Source: A. THz Channel Modeling · p. 4

B. THz Channel Conditions

THz channel behavior is shaped by molecular absorption and sparse propagation, with the relationship between range and array spacing separating operation into three regions. The regions have different channel-conditioning and spatial-multiplexing properties.

Molecular absorption

Operating regions

  • Region 1 has relatively large SA separation ∆ and an always well-conditioned channel, typical of a large-intelligent-surface setting.Source: B. THz Channel Conditions · p. 5
  • Region 2 can support efficient spatial multiplexing and spatial modulation on eigenchannels when the SA spacing ∆ is finely tuned, despite high channel correlation.This region occurs when communication distance D is very large compared with ∆.Source: B. THz Channel Conditions · p. 5
  • Beyond the Rayleigh distance, Region 2 optimizations no longer hold because antenna arrays have limited physical size; spatial oversampling is one proposed alternative, with potential noise-shaping benefits.The stated benefits include lower noise figures, increased linearity, and relaxed high-resolution data-converter requirements.Source: B. THz Channel Conditions · p. 5

IV. RECENT RESEARCH ADVANCES

THz signal and channel characteristics differ fundamentally from those of conventional systems, so established signal-processing problems must be revisited. The paper highlights challenges spanning beam control, channel knowledge, and data detection.

  • THz systems require renewed approaches to beamforming and beamsteering criteria, precoding and combining, channel estimation, and data detection.Source: §IV · p. 5
  • The highlighted challenges include accurate beam control and optimal precoding and combining methods.Source: §IV · p. 5
  • Low-cost channel-estimation paradigms and near-optimal data detection are also identified as relevant research problems.Source: §IV · p. 5

A. Modulation

The surveyed advances address THz hardware limits with pulse-based modulation and call for THz-specific waveform, resource-allocation, and beamforming designs. Frequency-dependent path directions also motivate new hybrid precoding and signal-processing approaches.

A. Modulation

  • Room-temperature graphene transceivers generate only very short, few-milliwatt pulses at higher THz frequencies, limiting long-distance communication; pulse-based asymmetric on-off keying is investigated.The cited scheme uses 100-femtosecond pulses for a logic one and can support up to a Tb/sec for many nano-devices.Source: A. Modulation · p. 5

B. Waveform Design and Beamforming

  • THz-specific multi-carrier waveform designs are needed to exploit spectral windows, while joint optimization can control power, sub-window allocation, and modulation formats.The channel is typically assumed flat and single-carrier, with possible carrier aggregation, although indoor multipath and frequency-selective receivers can introduce selectivity.Source: B. Waveform Design and Beamforming · p. 5
  • Frequency-dependent path directions cause beam squint and array-gain loss, motivating hybrid precoding and techniques such as sparse delay-Vandermonde factorizations [14].The losses are more significant in analog architectures.Source: B. Waveform Design and Beamforming · p. 5

C. Multi-Carrier Antenna Configurations

Interleaved antenna maps coordinate frequency use across densely packed AEs while preserving useful antenna spacing. Spatial and index modulation can encode information in activated antenna locations and frequency allocations, but high channel correlation makes spatial modulation challenging at very high frequencies.

C. Multi-Carrier Antenna Configurations

  • Interleaved antenna maps assign neighboring AEs to different absorption-transmission windows, while same-frequency AEs remain separated by δ = λ/2 and different-frequency AEs by δ = λSPP.Same-frequency SAs can also maintain optimal separation ∆ for multiplexing gains.Source: C. Multi-Carrier Antenna Configurations · p. 5, Figure 5 · p. 6

D. Spatial and Index Modulation

  • Spatial modulation encodes bits through activated antenna locations, but high channel correlation makes it challenging at very high frequencies.Adaptive and hierarchical schemes can map bits to antenna indices at either the SA or AE level.Source: D. Spatial and Index Modulation · pp. 5–6
  • Antenna maps can combine spatial modulation with frequency interleaving, assigning information bits to both antenna indices and frequency allocations.The AEs can be activated in real time to target a bit rate at a communication range.Source: D. Spatial and Index Modulation · p. 6

V. PROSPECTIVE RESEARCH DIRECTIONS

The paper identifies sensing and data centers as prospective settings for THz UM-MIMO, linking its capabilities to sensing applications and high-rate wireless links.

V. PROSPECTIVE RESEARCH DIRECTIONS

  • THz UM-MIMO is proposed for sensing, data centers, intelligent surfaces, and mobile mid-range communications.The paper presents these as prospective use cases.Source: §V · p. 6

A. Communications and Sensing

  • UM-MIMO could extend gas sensing distances by mitigating distance-dependent molecular absorption, while spatial degrees of freedom can improve temperature and displacement sensing resolution.The resulting sensing information can be communicated over a distributed wireless-sensing network.Source: A. Communications and Sensing · p. 6

B. Data Centers

  • Wireless data-center links could reduce costs and improve energy efficiency by eliminating power-hungry switches, but require traffic-aware channel scheduling and efficient networking.The paper also notes that THz absorption losses are significantly reduced in enclosed underwater data centers using nitrogen for cooling.Source: B. Data Centers · p. 6

C. Large Intelligent Surfaces

The paper discusses THz UM-MIMO as a basis for large intelligent surfaces and cell-free systems, while noting that coherent arrays are difficult to obtain at THz sizes.

C. Large Intelligent Surfaces

  • Large intelligent surfaces can use THz UM-MIMO, but very small THz antenna elements make coherent arrays difficult and make reconfigurable LISs a practical solution.The paper also describes active and passive LIS configurations, including graphene-based plasmonic metasurfaces with local tuning.Source: C. Large Intelligent Surfaces · pp. 6–7

D. Cell-Free Systems

  • Cell-free THz systems can place users near at least one of many densely deployed access points, using redundant paths to mitigate blockage and addressing THz range limits.The applicability of cell-free massive MIMO to THz remains an open problem.Source: D. Cell-Free Systems · p. 7

E. Mid-Range Mobile Communications

The paper presents THz UM-MIMO as a prospective option for mid-range mobile communications, while identifying blockage and multipath Doppler as challenges; its conclusion argues that suitable configurations can address distance and power limits.

E. Mid-Range Mobile Communications

  • Adaptive, compact THz UM-MIMO arrays can share transceiver resources, tune carrier frequencies, and direct beams to multiple users in mobile applications requiring several meters of coverage.The paper points to vehicular networks and drones as potential applications.Source: E. Mid-Range Mobile Communications · p. 7
  • Blockage from small wavelengths and suspended particles challenges mobile THz links, while multipath fading Doppler spread is more severe even though LoS Doppler is minimized.The paper proposes coexistence with mmWave and microwave systems; mature mmWave solutions may serve high-mobility links while THz handles backhaul.Source: E. Mid-Range Mobile Communications · p. 7

VI. CONCLUSION

  • The paper argues that properly configured UM-MIMO arrays can overcome distance and power limitations and identifies graphene-based nano-antenna arrays as a potential way to realize THz UM-MIMO.It also highlights modulation, waveform design, and resource allocation as research advances for THz communications.Source: §VI · p. 7
Loading 1902.11090v4…