PaperlayerView exact revision on arXiv

Source-linked AI summary

Scoring the Terabit/s Goal:Broadband Connectivity in 6G

Nandana Rajatheva, Italo Atzeni, Simon Bicais, Emil Bjornson, Andre Bourdoux, Stefano Buzzi, Carmen D'Andrea, Jean-Baptiste Dore, Serhat Erkucuk, Manuel Fuentes, Ke Guan, Yuzhou Hu, Xiaojing Huang, Jari Hulkkonen, Josep Miquel Jornet, Marcos Katz, Behrooz Makki, Rickard Nilsson, Erdal Panayirci, Khaled Rabie, Nuwanthika Rajapaksha, MohammadJavad Salehi, Hadi Sarieddeen, Shahriar Shahabuddin, Tommy Svensson, Oskari Tervo, Antti Tolli, Qingqing Wu, Wen Xu

arXiv:2008.07220v2eess.SP

TL;DR

The paper addresses how 6G can provide broadband connectivity at rates up to the Tbps range across diverse use cases. It reviews PHY/MAC technologies and their challenges, reporting cell-free rate advantages for most users and IAB coverage approaching fiber-connected performance under specified settings.

Abstract

from arXiv · show

This paper explores the road to vastly improving the broadband connectivity in future 6G wireless systems. Different categories of use cases are considered, with peak data rates up to 1 Tbps. Several categories of enablers at the infrastructure, spectrum, and protocol/algorithmic levels are required to realize the intended broadband connectivity goals in 6G. At the infrastructure level, we consider ultra-massive MIMO technology (possibly implemented using holographic radio), intelligent reflecting surfaces, user-centric cell-free networking, integrated access and backhaul, and integrated space and terrestrial networks. At the spectrum level, the network must seamlessly utilize sub-6 GHz bands for coverage and spatial multiplexing of many devices, while higher bands will be mainly used for pushing the peak rates of point-to-point links. Finally, at the protocol/algorithmic level, the enablers include improved coding, modulation, and waveforms to achieve lower latency, higher reliability, and reduced complexity.

I. INTRODUCTION

The paper frames 6G broadband as a path toward Tbps-range wireless rates amid questions about coverage and quality of service. It reviews relevant 5G foundations and organizes proposed 6G enablers across spectrum, infrastructure, and protocol/algorithmic levels.

I. INTRODUCTION

  • 6G is framed around questions of what technologies will define it, whether 1 Tbps is practical, and whether uniform coverage and quality of service can be achieved.The paper also describes possible infrastructure densification, higher-frequency links, and integration of terrestrial and satellite networks.Source: §I · p. 2

A. RELATED WORKS

  • Prior 6G visions emphasize varied priorities, including battery life, flexibility, multidisciplinary design, sub-THz and optical communications, and high-quality wireless access.Related work also surveys spectrum, PHY/MAC, architecture, intelligence, and large distributed antenna arrays.Source: A. RELATED WORKS · pp. 2–3

B. PAPER CONTRIBUTION AND ORGANIZATION

  • The paper focuses on PHY- and MAC-layer technologies and research challenges for broadband connections reaching 1 Tbps in multiuser scenarios.It distinguishes this focus from general 6G overviews and emphasizes technical depth, equations, and illustrative numerical results.Source: B. PAPER CONTRIBUTION AND ORGANIZATION · p. 3
  • The paper groups Tbps-connectivity enablers into spectrum, infrastructure, and protocol/algorithmic categories, covering THz and optical bands, cell-free networking, IAB, and new coding and modulation approaches.It also discusses ultra-massive MIMO, holographic radio, intelligent reflecting surfaces, non-terrestrial networks, machine learning, and coded caching.Source: B. PAPER CONTRIBUTION AND ORGANIZATION · pp. 3–4

II. FROM 5G TO 6G

  • The paper frames its 6G use-case vision with a state-of-the-art 5G overview: 5G NR foundations include OFDM/DFT-s-OFDM waveforms, LDPC and polar coding, QPSK/QAM modulation, MIMO, and Release 16 IAB.Source: §II · p. 4, A. 5G TODAY · pp. 4–6

B. 6G USE CASES

The paper considers 6G broadband applications with high data-rate and availability demands alongside specialized requirements. These use cases range from xHaul and mobile hotspots to industrial control, autonomous mobility, extended reality, and remote-area access.

B. 6G USE CASES

  • 6G applications are expected to require broadband connectivity with high data rates and availability, while differing in their specialized performance requirements.The use-case list complements rather than replaces existing 5G use cases.Source: B. 6G USE CASES · p. 6
  • Extreme-capacity xHaul uses fixed symmetric point-to-point links that combine high bandwidth and spectral efficiency to aggregate user traffic and complement limited fiber deployment.The links can also provide access to edge or cloud computing resources.Source: B. 6G USE CASES · p. 6
  • High-mobility hotspots need high data rates, low latency, and near-100% availability at speeds of hundreds of km/h, with several GHz of bandwidth supporting up to 100 Gbps.The mobile-hotspot-to-fixed-infrastructure link is identified as the biggest challenge in this scenario.Source: B. 6G USE CASES · pp. 6–7
  • Multi-sensory extended reality requires per-user rates in the Gbps range and very low latency, while multi-view holographic communications may require data rates on the order of terabits per second.Remote areas are another use case: 6G targets 10 Mbps in every populated area using ground-based and spaceborne network components, with cost-efficient deployment.Source: B. 6G USE CASES · p. 7
  • Industrial automation targets reliability of 1-10^-9 and round-trip latency around 0.1-1 ms, while connected autonomous vehicles require reliability above 99.99999% and latency below 1 ms.Autonomous mobility scenarios also include drone-delivery systems and swarms requiring improved capacity and coverage.Source: B. 6G USE CASES · p. 7

C. 6G KEY PERFORMANCE INDICATORS

The paper frames 6G as requiring sharper and interdependent KPIs, with configurable KPI groups for specialized applications. It illustrates routes to 1 Tb/s through bandwidth and spectral efficiency or aggressive spatial multiplexing.

6G KPI framework

  • 6G is expected to retain many existing KPIs while making minimum requirements orders of magnitude sharper and considering new societal, sustainability, and productivity indicators.Unlike largely independent 5G KPIs, 6G KPIs are envisaged in groups whose indicators must be met together.Source: C. 6G KEY PERFORMANCE INDICATORS · p. 8
  • Because specialized application requirements cannot all be achieved together, 6G will need real-time configuration for different KPI groups.Source: C. 6G KEY PERFORMANCE INDICATORS · p. 8

Envisaged KPIs

  • The paper envisages peak data rates up to 1 Tb/s, peak spectral efficiency up to 60 b/s/Hz, bandwidth above 100 GHz in sub-THz, THz, and visible-light bands, and latency down to 0.1 ms.Other envisaged targets include 1 Gb/s user-experienced rates, 3 b/s/Hz user-experienced spectral efficiency, 1 terabit per Joule, and jitter down to 1 µs.Source: C. 6G KEY PERFORMANCE INDICATORS · pp. 8–9

Scoring the Terabit/s goal

  • With 100 GHz, 10 b/s/Hz is sufficient for 1 Tb/s; two parallel 16-QAM streams can achieve that spectral efficiency.The paper notes that peak KPI values are unlikely to be achieved simultaneously.Source: C. 6G KEY PERFORMANCE INDICATORS · p. 9
  • With 1 GHz and 10 b/s/Hz, spatially multiplexing 100 devices reaches an aggregate capacity of 1 Tb/s, and more aggressive multiplexing could do so below 6 GHz.Source: C. 6G KEY PERFORMANCE INDICATORS · p. 9

III. ENABLERS AT THE SPECTRUM LEVEL

The spectrum vision combines existing and higher-frequency bands through a joint interface and seamless handover. THz links offer very wide bandwidths, but their propagation, hardware, channel characterization, and networking constraints require specialized enablers.

III. ENABLERS AT THE SPECTRUM LEVEL

  • 6G is envisioned to dynamically allocate users across bands, with lower frequencies often favoring reliability, mobility, and connectivity density and higher frequencies favoring peak rates and latency.The vision is a converged interface with seamless handover, not a gradual shift to progressively higher operating frequencies.Source: §III · p. 10

A. ENABLERS FOR TERAHERTZ COMMUNICATIONS

  • THz systems can support rates above 1 Tbps through larger bandwidths and hundreds to thousands of spectrum-sharing users through shorter ranges.Source: A. ENABLERS FOR TERAHERTZ COMMUNICATIONS · p. 10
  • THz propagation combines spreading and molecular absorption losses; absorption creates distance- and medium-dependent transmission windows ranging from tens to hundreds of GHz.High-gain directional antennas are needed to compensate for spreading loss and the limited transmission power of THz sources.Source: A. ENABLERS FOR TERAHERTZ COMMUNICATIONS · pp. 10–11
  • THz channel sounding is costly and time-consuming, so sparse measurements may be extended using measurement-calibrated ray-tracing simulations.THz directionality and attenuation make channel sounding more challenging than at traditional cellular frequencies.Source: A. ENABLERS FOR TERAHERTZ COMMUNICATIONS · p. 12
  • At 300–308 GHz, measurements in an HST wagon agreed with simulated power-delay profiles on significant paths in power and delay.The measurement used an 8 GHz bandwidth; the measurement-validated ray-tracing simulator can generate channel data for other setups.Source: A. ENABLERS FOR TERAHERTZ COMMUNICATIONS · p. 12, Figure 6 · p. 12, Figure 7 · p. 12
  • THz links face distance challenges from high spreading loss and low transmitter power, while highly directional antennas require revised access, discovery, and relaying strategies.Receiver-initiated polling and use of the full antenna radiation pattern for neighbor discovery are cited as proposed approaches.Source: A. ENABLERS FOR TERAHERTZ COMMUNICATIONS · p. 13
  • THz basebands need efficient, parallelizable coding and signal processing because channels can exceed 100 GHz while state-of-the-art converters sample at about 100 Gigasamples per second.The paper calls for co-optimization of channel estimation, coding, and data detection, and consideration of low-resolution conversion and all-analog solutions.Source: A. ENABLERS FOR TERAHERTZ COMMUNICATIONS · p. 13

B. ENABLERS FOR OPTICAL WIRELESS

Optical wireless communication (OWC) is presented as a complement to radio, with potential for high-speed, low-power, and secure access and backhaul. Its practical use depends on addressing alignment, receiver sensitivity, channel characterization, and network-design challenges.

OWC overview

  • OWC can complement RF or wired links for access and backhaul, with potential for ubiquitous, high-speed, low-power, secure, and low-cost wireless access.The paper describes hybrid OWC networks for IoT connectivity and flexible backhaul/fronthaul links.Source: COMMUNICATIONS · p. 13
  • OWC spans infrared, visible-light, and ultraviolet bands; FSO commonly denotes long-range outdoor or space laser links, while VLC uses LEDs, mostly indoors and in vehicles.Source: COMMUNICATIONS · pp. 13–14
  • An FSO system achieved 13.16 Tbps over 10 km, while indoor VLC setups have reached multiple Gbps.The paper cautions that VLC energy-efficiency comparisons with RF are difficult because VLC also provides illumination.Source: COMMUNICATIONS · p. 14

Challenges and research directions

  • OWC faces transmitter-receiver alignment requirements on long-range links and receive-sensitivity limits from existing detector technology.Source: COMMUNICATIONS · p. 14

Visible light communications

  • VLC supports bidirectional data delivery through lighting, but its positive-valued intensity-modulated signal constraint and blocked line of sight motivate hybrid optical-radio networks.The paper describes horizontal handover between light sources and vertical handover to RF.Source: COMMUNICATIONS · p. 14

Challenges and research directions

  • Open OWC research includes channel models that account for user mobility and orientation, low-complexity multi-user access, and cellular VLC architectures with scheduling and handover.The stated goals include capacity, low latency, and fairness.Source: COMMUNICATIONS · p. 15
  • Dense OWC IoT and medium-range links raise network-topology, MAC, environmental-condition, and multi-hop design challenges.The paper notes that multi-hop transmission can reduce delay for vehicle safety information.Source: COMMUNICATIONS · p. 15

IV. ENABLERS AT THE INFRASTRUCTURE LEVEL

Multiplexing gain can be increased by serving more UEs simultaneously, including through larger or distributed antenna arrays.

IV. ENABLERS AT THE INFRASTRUCTURE LEVEL

  • Larger and distributed antenna arrays are among the technologies for increasing the number of simultaneously active UEs in 6G.Source: §IV · p. 15

A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC RADIO

The section argues that 6G should move beyond 5G’s restricted beam-space massive MIMO toward ultra-massive arrays that exploit more spatial dimensions. It presents holographic radio as a continuous-aperture approach, while noting substantial modeling and signal-processing challenges.

A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC RADIO

  • Ultra-massive MIMO extends spatial multiplexing with large antenna arrays, while 5G beam-space designs restrict users to predefined beams and work best under limited channel conditions.The 2D DFT-based approximation is optimal only without near-field scattering, with predefined array geometries, perfect calibration, and no mutual coupling.Source: A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC… · pp. 15–16
  • Using more spatial dimensions matters especially in scarce sub-6 GHz bands, where longer coherence time can accommodate channel estimation for more multiplexed users.The passage states that 1 MHz at 1 GHz can give the same data rate as 100 MHz at 100 GHz because the lower band supports 100 times more users.Source: A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC… · p. 16
  • Fully digital arrays with hundreds or thousands of phase-synchronized antennas could use more spatial dimensions, although implementation complexity grows with carrier frequency.The section also anticipates doubly massive MIMO links at mmWave frequencies, with many antennas at both ends.Source: A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC… · p. 16
  • Continuous-aperture antennas can improve beamforming accuracy with cleaner transmissions and smaller sidelobes, but dense conventional arrays would make individual RF chains costly and energy-intensive.Holographic radio is presented as a possible active continuous-aperture implementation.Source: A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC… · p. 16
  • Holographic radio records and reconstructs electromagnetic fields through coherent-wave interference, enabling pixel-level, high-density spatial multiplexing rather than sparse 5G beam-space.The section describes spherical-wave generation across the array and holographic techniques for obtaining channel information and synthesizing downlink fields.Source: A. ULTRA-MASSIVE MIMO AND HOLOGRAPHIC… · p. 16, Realization of holographic radio · p. 17
  • Holographic radio’s performance estimation requires detailed electromagnetic computation, and its physical-layer algorithms must adapt to optical-domain processing.The section identifies missing integrated communication-and-electromagnetic models, hardware mapping, photonics integration, and optical computing as open challenges.Source: Realization of holographic radio · p. 17, Signal processing for holographic radio · p. 17

B. INTELLIGENT REFLECTING SURFACES

Intelligent reflecting surfaces (IRSs) add controllable scattering to wireless links by tuning how surface elements reflect signals. The section describes their operating principles, deployment considerations, potential uses, and unresolved practical constraints.

B. INTELLIGENT REFLECTING SURFACES

  • An IRS uses controllable surface elements to shape reflected signals, creating a programmable propagation environment without conventional RF chains.It is a thin, large-area surface that can be integrated on building facades, walls, or ceilings; its elements adjust scattering properties such as phase delay, amplitude, and polarization.Source: B. INTELLIGENT REFLECTING SURFACES · p. 18
  • Tuning each element’s phase delay can align reflected paths with the direct path, while joint optimization with transceivers can target link spectral or energy efficiency.A specular-reflector configuration is generally suboptimal; an IRS can also focus the reflected waveform at a receiver’s location.Source: B. INTELLIGENT REFLECTING SURFACES · pp. 18–19
  • An IRS greatly improves spectral efficiency when the direct path is weak, while gains are more modest when that path is strong.The illustrated single-antenna example needs hundreds of elements for effectiveness, although more than 1000 sub-wavelength elements can fit on a 1 m × 1 m surface.Source: Figure 10 · p. 19, B. INTELLIGENT REFLECTING SURFACES · p. 19
  • IRSs are presented as promising for short-range sub-THz and THz links, where they can add a strong path when line of sight is blocked.They do not amplify the incident signal, and their physically large surfaces and beam-squinting constrain comparisons with regenerative relays.Source: B. INTELLIGENT REFLECTING SURFACES · pp. 17, 19
  • For a 45 m transmitter–receiver link with a 1024-element IRS, channel gain improves most when the surface is close to either endpoint.The section proposes placing a large IRS near a small access point or in a room with good channels to both access point and user.Source: B. INTELLIGENT REFLECTING SURFACES · pp. 19–20, Figure 12 · p. 20
  • Beyond strengthening a single-user signal, an IRS can improve channel rank, suppress interference, and enhance multicasting.The section lists further prospective uses including wireless power transfer, physical-layer security, coordinated multipoint transmission, UAV links, and backscattering.Source: B. INTELLIGENT REFLECTING SURFACES · p. 20
  • Practical IRS control remains constrained by limited phase resolution, hardware effects omitted from existing models, broadband phase trade-offs, and difficult channel estimation.The section calls for experimentally validated models; estimation and control resources grow linearly with element count, and passive surfaces cannot send or receive pilots.Source: B. INTELLIGENT REFLECTING SURFACES · p. 20

C. SCALABLE CELL-FREE NETWORKING WITH USER-CENTRIC ASSOCIATION

Cell-free massive MIMO replaces co-located arrays with many cooperating, distributed APs, aiming for more uniform service across UE locations. A user-centric association limits each UE to selected APs, balancing performance with signaling and backhaul overhead.

Cell-free architecture

  • Distributed APs jointly serve UEs through CPU-connected fronthaul, placing each UE near favorable antennas and improving diversity against blockage and large-scale fading.The architecture replaces co-located massive arrays with many low-cost APs equipped with few antennas.Source: C. SCALABLE CELL-FREE NETWORKING WITH… · p. 21, Figure 13 · p. 22

User-centric association

  • Each UE can be served by a limited number of APs selected using large-scale fading, while FCF has every AP serve every UE.The described simple association selects a predetermined number of APs with the strongest average channel conditions.Source: C. SCALABLE CELL-FREE NETWORKING WITH… · pp. 21–22

Performance comparison

  • About 90% of downlink UEs and 80% of uplink UEs obtain much better rates with FCF and CF-UC than with mMIMO; mMIMO leads only for the 90%-likely per-user rate.The comparison uses MR processing and a heavily loaded setup with 30 UEs; the CDFs and tables report the comparison.Source: C. SCALABLE CELL-FREE NETWORKING WITH… · p. 23, Figure 14 · p. 24, Table 4 · p. 24
  • The mMIMO comparison is specific to simple MR processing and a heavily loaded coherence block; lower load and regularized zero-forcing reduce its gap with cell-free deployments.The authors caution against generalizing the comparison as evidence that massive MIMO does not work.Source: C. SCALABLE CELL-FREE NETWORKING WITH… · p. 23
  • CF-UC with NUC ≥10 provides performance comparable to or better than FCF, with less signaling and backhaul overhead; performance depends crucially on UE-AP association.FCF corresponds to CF-UC with NUC = M, and the results indicate an optimal NUC value for system performance.Source: C. SCALABLE CELL-FREE NETWORKING WITH… · p. 23, Figure 15 · p. 24

Initial access and implementation challenges

  • Cell-free initial access requires redesigned cell identification and random access so multiple APs can support a UE, while greater centralized coordination can raise rates at the cost of complexity, latency, and fronthaul signaling.Cell search and random access affect latency, energy consumption, and supported UE count; joint CSI processing at the CPU adds coordination overhead.Source: C. SCALABLE CELL-FREE NETWORKING WITH… · pp. 23, 25

D. INTEGRATED ACCESS AND BACKHAUL

Integrated access and backhaul (IAB) offers wireless backhaul as a flexible alternative to fiber for dense networks, where fiber installation can be costly or delayed. The analysis finds that IAB can approach fiber-connected coverage with a modest increase in SBS density, while offering deployment flexibility and potential cost and time savings.

Motivation

  • Fiber offers reliable transport at demonstrated Tbps rates, but its installation can require substantial investment, take a long time, or be disallowed in some metropolitan areas.Microwave backhaul supports up to 100 Gbps and is described as a lower-cost, more flexible deployment option,,.Source: D. INTEGRATED ACCESS AND BACKHAUL · p. 25

Performance

  • The analysis models a two-tier HetNet with wireless access and backhaul at SBSs, fiber-connected IAB donors, random node placement, and blockage-aware mmWave propagation.Its blockage model accounts for correlated obstacle-induced blocking [227, Chapter 14].Source: D. INTEGRATED ACCESS AND BACKHAUL · p. 26
  • At a 100 Mbps target rate, an IAB density of 85 km^-2 matches the 0.81 coverage probability of a fully fiber-connected network at 60 SBSs/km^-2; with 30% fiber-connected SBSs, 70 km^-2 suffices.The authors say these FHPPP-based results are pessimistic and that the relative IAB-to-fiber performance gap decreases as network density rises.Source: D. INTEGRATED ACCESS AND BACKHAUL · p. 27, Figure 18 · p. 27, Figure 19 · p. 27
  • IAB significantly improves coverage over an MBS-only network, and its relative performance gap from fiber-connected networks decreases as network density increases.The comparison uses coverage probability, defined as the probability that UEs meet their minimum target rates.Source: D. INTEGRATED ACCESS AND BACKHAUL · p. 27, Figure 19 · p. 27

Benefits

  • IAB nodes can be placed wherever they have a fairly good connection to parent nodes, increasing network flexibility and the scope for topology optimization.The paper also gives urban estimates of 20,000 GBP/km for fiber versus around 2,500 GBP per SBS [228, Table 7],.Source: D. INTEGRATED ACCESS AND BACKHAUL · p. 27
  • IAB is described as robust to blockage, rain, and tree foliage with proper antenna height and network density, except in suburban areas with moderate or high tree foliage.Fiber installation delays also make a rapid IAB deployment followed by possible fiber replacement an expected setup.Source: D. INTEGRATED ACCESS AND BACKHAUL · pp. 27–28

E. INTEGRATED SPACE AND TERRESTRIAL

Integrated space and terrestrial networks (ISTNs) combine spaceborne, airborne, and ground-based layers to support global high-speed communications. Their capacity is constrained by aerial-backbone bandwidth and air-to-ground spectral efficiency, while mmWave links and LoS-MIMO are presented as potential enablers with important current limitations.

NETWORKS

  • ISTNs combine spaceborne, airborne, and ground-based network layers, using satellites, aerial platforms, and terrestrial infrastructure for global communications.The architecture is illustrated in Fig. 20; its lines distinguish free-space optics, mmWave, and other microwave links.Source: NETWORKS · p. 28, Figure 20 · p. 29
  • The main capacity bottlenecks are limited bandwidth for high-speed aerial backbones and low spectral efficiency on direct air-to-ground links.Microwave-based aerial backbones have limited bandwidth, while satellite beams cover large ground areas and limit frequency reuse.Source: NETWORKS · p. 28
  • An mmWave backbone using more than 10 GHz of E-band bandwidth is proposed for rates up to 100 Gbps, with LoS-MIMO available for multi-channel transmission.The cited example estimates about 10 dB of atmospheric and cumulus loss over a 50 km link.Source: NETWORKS · p. 28
  • Current mmWave links still fall short of high-capacity ISTN needs: full-duplex cancellation is difficult, LoS-MIMO channel rank is limited, and links beyond 50 km are difficult even at 40 dBm.The paper points to massive, preferably conformal and reconfigurable, antenna arrays and lists channel modeling, 3D networking, and protocol optimization as other enablers.Source: NETWORKS · pp. 28–29

F. INTEGRATED WIDEBAND BROADCAST NETWORKS

Broadcast and multicast can serve many users requesting the same content, including where access points have limited beamforming capability. The section discusses combining mobile broadband and broadcast networks and reports power and capacity benefits for a proposed 6G wideband broadcasting system.

F. INTEGRATED WIDEBAND BROADCAST NETWORKS

  • Broadcast and multicast provide simultaneous delivery of the same content to many devices with predefined quality of service, and can offer scalable delivery when AP beamforming is limited.An AP able to send only wide beams can broadcast information across the beam's coverage area.Source: F. INTEGRATED WIDEBAND BROADCAST… · p. 29
  • The section presents convergence of mobile broadband and broadcast as a way to combine low-power cellular coverage with complementary high-power broadcast coverage.Potential approaches include a shared radio physical layer, common transport protocols, or applications that adapt to underlying network capabilities.Source: F. INTEGRATED WIDEBAND BROADCAST… · p. 29, Figure 21 · p. 30
  • The proposed 6G wideband broadcasting system uses frequency reuse-1 and is reported to save around 90% of total transmit power while increasing capacity by 37–60% for the same coverage.It uses about 17 dB less power per 8 MHz RF channel and half the transmission power of a single traditional digital RF channel.Source: F. INTEGRATED WIDEBAND BROADCAST… · p. 30, Figure 22 · p. 30

V. ENABLERS AT THE PROTOCOL/ALGORITHMIC

Spectral efficiency is linked to modulation and coding, interference management, and resource management.

V. ENABLERS AT THE PROTOCOL/ALGORITHMIC

  • Modulation and coding are identified as contributors to spectral efficiency.The section identifies spectral efficiency as the third factor in equation (1).Source: §V · p. 30
  • Interference management is identified as a contributor to spectral efficiency.The section identifies spectral efficiency as the third factor in equation (1).Source: §V · p. 30
  • Resource management is identified as a contributor to spectral efficiency.The section identifies spectral efficiency as the third factor in equation (1).Source: §V · p. 30

A. CODING, MODULATION, WAVEFORM, AND DUPLEX

The section surveys coding, modulation, waveform, duplex, interference-management, and THz spatial-tuning approaches relevant to 6G broadband links. It emphasizes trade-offs between performance, throughput, energy use, robustness, and implementation complexity.

Channel coding

  • 6G coding must raise decoder throughput and energy efficiency while balancing error performance, code length, parallelization, and iterations.The required energy consumption per bit is expected to fall by 1–2 orders of magnitude to maintain current devices’ energy consumption.Source: A. CODING, MODULATION, WAVEFORM, AND… · p. 30
  • Polar-code belief propagation offers a more parallelizable alternative to serial CA-SCL decoding, while permuted factor graphs can reduce iterations between permutations.The partially permuted graph retains information about variable nodes in unpermuted subgraphs.Source: A. CODING, MODULATION, WAVEFORM, AND… · pp. 30–31, Figure 23 · p. 31

Modulation and waveform

  • No single waveform suits every 6G scenario: single-carrier waveforms may suit high-frequency links, while OFDM variants remain an option for indoor hotspots.The cited trade-offs include phase noise, propagation loss, power-amplifier efficiency, data rates, and flexible scheduling.Source: A. CODING, MODULATION, WAVEFORM, AND… · p. 31
  • Under strong phase noise, Polar-QAM reaches larger spectral efficiencies than QAM, while medium phase noise brings a minor performance loss versus QAM.Its adjustable constellation shape trades increased PAPR for greater phase-noise robustness.Source: A. CODING, MODULATION, WAVEFORM, AND… · p. 32, Figure 25 · p. 34

Full-duplex

  • In-band full duplex can potentially increase throughput, but self-interference requires passive suppression, analog cancellation, and digital cancellation.The technique remains an open 6G research area, with challenges spanning antenna and circuit design through network theory.Source: A. CODING, MODULATION, WAVEFORM, AND… · pp. 32–33

Protocol-level interference management

  • NOMA and rate splitting offer protocol-level alternatives when massive-MIMO interference management is insufficient, but their coupled decoding and resource decisions increase complexity.NOMA fully decodes and removes multi-user interference; rate splitting partially decodes it and partially treats it as noise.Source: A. CODING, MODULATION, WAVEFORM, AND… · pp. 33–34

THz-band spatial tuning

  • In THz links, spatial tuning adjusts antenna separation to preserve channel rank, but becomes ineffective beyond the achievable Rayleigh distance.Allocating 1000 antenna elements per subarray yielded a 60 dB performance enhancement in the illustrated LoS case; exact channel characterization still requires relaxing assumptions such as uniform arrays and LoS propagation.Source: A. CODING, MODULATION, WAVEFORM, AND… · pp. 34–35, Figure 27 · p. 35

B. MACHINE LEARNING-AIDED ALGORITHMS

The section considers machine learning as a complement to model-driven communication algorithms when models are deficient or conventional algorithms are complex. It reviews applications in end-to-end system design, channel estimation and detection, and resource management, alongside training and deployment challenges.

Motivation

  • Machine learning may help when channel properties are unknown or conventional signal-processing algorithms are too complex for latency-critical or jointly optimized tasks.The paper describes these as modeling and algorithmic deficiencies, respectively.Source: B. MACHINE LEARNING-AIDED ALGORITHMS · pp. 35–36

End-to-end learning

Joint channel estimation and detection

  • Learning-based joint channel estimation and detection can reduce signaling overhead and handle nonlinear clipping or unknown hardware-impairment models.Online learning has also been proposed to jointly handle fading channels and nonlinear distortion.Source: B. MACHINE LEARNING-AIDED ALGORITHMS · p. 36

Resource management

  • Machine learning is applied to resource allocation and hybrid precoding, including approaches reported to reduce complexity versus exhaustive-search optimization.The section also identifies offline training mismatch as a challenge and points to online training and realistic training data as future directions.Source: B. MACHINE LEARNING-AIDED ALGORITHMS · pp. 36–37

C. CODED CACHING

Coded caching (CC) uses shared cache memories and multicast codewords to increase delivery rates, with additional gains possible alongside multi-antenna spatial multiplexing. Its practical performance depends on subpacketization, beamforming, SNR, and the caching overhead.

C. CODED CACHING

  • CC creates a global caching gain by multicasting codewords that carry useful data for multiple users, in addition to each user's local caching gain.With K users, cache size M per user, and N files, the passage describes the CC gain as proportional to the total cache size across users.Source: C. CODED CACHING · p. 37
  • Multi-antenna CC can add its caching gain to spatial multiplexing gain by beamforming parallel multicast codewords to user groups and suppressing inter-codeword interference.The number of parallel multicast messages can vary with available spatial degrees of freedom and power.Source: C. CODED CACHING · p. 37
  • Multi-antenna CC can reduce subpacketization growth to linear or near-linear when spatial multiplexing gain exceeds CC gain, addressing a major implementation problem.Original single- and multi-antenna CC schemes have exponentially growing subpacketization requirements.Source: C. CODED CACHING · p. 37
  • Immersive applications offer caching opportunities because viewpoint content is partly shared across users, despite being location- and time-dependent.User-device caching and edge computation offloading could support high-throughput, low-latency traffic with stability and reliability for this use case.Source: C. CODED CACHING · pp. 37–38, Figure 30 · p. 38
  • Poor channel conditions can degrade CC performance, while proactive cache placement adds communication overhead and energy consumption that can reduce energy efficiency.The literature's treatment of CC energy efficiency is described as largely unaddressed.Source: C. CODED CACHING · p. 38
  • Below 15 dB, RED with optimized beamforming outperforms MS with zero-forcing, while that comparison reverses above 15 dB; the beamformer gap vanishes at high SNR.MS generally performs better than RED, and optimized beamforming generally outperforms zero-forcing, but MS requires exponentially growing subpacketization and optimized beamforming requires computationally intensive non-convex optimization.Source: Figure 31 · p. 38, C. CODED CACHING · p. 38

VI. 6G BROADBAND CONNECTIVITY FOR RURAL/UNDERDEVELOPED AREAS

Rural and underdeveloped areas remain underserved by successive wireless generations, widening the digital divide. The section discusses massive MIMO, cell-free systems, satellites, and intelligent reflecting surfaces as ways to address coverage, while describing practical trade-offs and constraints.

VI. 6G BROADBAND CONNECTIVITY FOR RURAL/UNDERDEVELOPED AREAS

  • Almost 4 billion people still lack broadband connectivity, and past wireless generations have generally delivered less benefit in rural areas than in population centers.The passage says this imbalance has sometimes resulted in worse rural coverage than previous generations and has increased the digital divide.Source: §VI · p. 39
  • Cell-free massive MIMO with max-min power control offers higher energy efficiency and throughput per user in rural environments than a single-cell system.Distributed antennas reduce path loss, and only a few antennas need to transmit at full power; conventional massive MIMO's cell-edge gains can come at a reduced sum-rate.Source: §VI · p. 39
  • Long-range massive MIMO and coherent transmission between adjacent access points can extend coverage, but rural channels face shadowing, blocking, and fading challenges.Coherent transmission in distributed cell-free arrangements also requires substantial backhaul or fronthaul capacity, and synchronization is difficult with large propagation-delay variations.Source: §VI · p. 39
  • LEO satellites and HAPS can provide broad rural coverage, fill terrestrial coverage gaps, and supply backhaul, but LEO deployment is costly and coordination faces power, interference, and synchronization constraints.The passage notes that many LEO satellites are needed for constant rural availability, and that coherent multi-satellite or space-ground transmission needs higher fronthaul capacity.Source: §VI · p. 39
  • IRSs can reflect signals around blocked line-of-sight paths in remote terrain without requiring backhaul or grid power, but long links need larger surfaces and remote control may limit mobility.The passage also notes exposure to harsh weather and sabotage for IRSs deployed in nature.Source: §VI · p. 39

VII. CONCLUDING REMARKS

The paper surveys PHY- and MAC-layer candidate technologies for Tbps wireless connectivity and organizes them as spectrum, infrastructure, and protocol/algorithmic enablers. Its vision prioritizes consistently high user rates, using different technologies across bands and integrating satellite and terrestrial networks to reach populations outside urban centers.

VII. CONCLUDING REMARKS

  • The survey categorizes candidate PHY- and MAC-layer technologies for Tbps 6G connectivity into spectrum, infrastructure, and protocol/algorithmic enablers.The paper also summarizes challenges, potential solutions, and open research questions in Table 6.Source: §VII · p. 39, Table 6 · p. 40
  • The vision shifts from network-centric extreme peak rates in special cases to user-centric prioritization of consistently high rates.It identifies cell-free massive MIMO and integrated access and backhaul, complemented by IRSs, as means of delivering ubiquitous connectivity.Source: §VII · p. 40
  • Sub-6 GHz is envisioned for wide-area coverage and spatial multiplexing, while sub-THz and THz bands target short-range rates near 1 Tbps.The paper says novel coding, modulation, and waveforms are needed for extreme rates with manageable complexity and robustness to hardware impairments; satellite-terrestrial integration is presented as a way to reach populations outside urban centers.Source: §VII · p. 40

Related papers

Loading 2008.07220v2…