PaperlayerView exact revision on arXiv

Source-linked AI summary

6G Wireless Systems: Vision, Requirements, Challenges, Insights, and Opportunities

Harsh Tataria, Mansoor Shafi, Andreas F. Molisch, Mischa Dohler, Henrik Sjöland, Fredrik Tufvesson

arXiv:2008.03213v2eess.SPcs.IT

TL;DR

Existing 6G research lacks a holistic top-down account of use cases, requirements, and system challenges. This paper synthesizes these layers and distinguishes what may be achievable over the next decade from what is possible.

  • Problem

    Existing 6G studies lack a holistic top-down account spanning use cases, technical requirements, and research challenges.

    Source: §VIII · p. 26

  • Method

    The paper surveys 6G use cases and requirements, then discusses deployment scenarios, research challenges, and possible solutions from applications through core networks to PHY.

    Source: §VIII · p. 26

  • Results

    The paper distinguishes what 6G may be achievable over the next decade from what is possible.

    Source: §VIII · p. 26

  • Takeaways & Limitations

    Supporting emerging 6G applications calls for bandwidth beyond 5G and new PHY and higher-layer capabilities.

    Source: §I · p. 1

    At THz bands, diffuse-scattering behavior lacks validation through measurements.

    Source: A. MmWave and THz Propagation Channels · p. 19

Abstract

from arXiv · show

Mobile communications have been undergoing a generational change every ten years or so. However, the time difference between the so-called "G's" is also decreasing. While fifth-generation (5G) systems are becoming a commercial reality, there is already significant interest in systems beyond 5G, which we refer to as the sixth-generation (6G) of wireless systems. In contrast to the already published papers on the topic, we take a top-down approach to 6G. We present a holistic discussion of 6G systems beginning with lifestyle and societal changes driving the need for next generation networks. This is followed by a discussion into the technical requirements needed to enable 6G applications, based on which we dissect key challenges, as well as possibilities for practically realizable system solutions across all layers of the Open Systems Interconnection stack. Since many of the 6G applications will need access to an order-of-magnitude more spectrum, utilization of frequencies between 100 GHz and 1 THz becomes of paramount importance. As such, the 6G eco-system will feature a diverse range of frequency bands, ranging from below 6 GHz up to 1 THz. We comprehensively characterize the limitations that must be overcome to realize working systems in these bands; and provide a unique perspective on the physical, as well as higher layer challenges relating to the design of next generation core networks, new modulation and coding methods, novel multiple access techniques, antenna arrays, wave propagation, radio-frequency transceiver design, as well as real-time signal processing. We rigorously discuss the fundamental changes required in the core networks of the future that serves as a major source of latency for time-sensitive applications. While evaluating the strengths and weaknesses of key 6G technologies, we differentiate what may be achievable over the next decade, relative to what is possible.

I. INTRODUCTION … 2) Connectivity for All Things:

The paper frames 6G requirements around changing lifestyles and societal needs, connecting holographic and pervasive communications to demands for greater bandwidth and new physical- and higher-layer capabilities.

I. INTRODUCTION

  • Future applications beyond 5G, including holographic and immersive communications and mission-critical connectivity, require larger bandwidths and new physical- and higher-layer capabilities.The introduction positions 6G research as spanning use cases, requirements, and potential solutions, especially at the physical layer.Source: §I · p. 1

A. Drivers for 6G Systems: Lifestyle and Societal Changes

  • The paper identifies high-fidelity holography, connectivity for all things, and time-sensitive or time-engineered applications as three key lifestyle and societal drivers of 6G.It connects these changes to the design and outlook of future wireless networks.Source: A. Drivers for 6G Systems: Lifestyle… · p. 1

1) High Fidelity Holographic Society:

  • 4K video requires 15.4 Mbps per UE, while video viewing and demand for such content are expected to grow as communication increasingly moves toward augmented reality.The paper also describes video communication supporting remote work, education, and events during the COVID-19 pandemic, with some developments expected to persist afterward.Source: 1) High Fidelity Holographic Society · pp. 1–2
  • A tiled 6’×20” human hologram could require 4.32 Tbps, with multiple viewpoints and synchronization needed for seamless delivery.Holographic telepresence could support remote troubleshooting, surgery, and education, but its transmission rates substantially exceed 5G capabilities.Source: 1) High Fidelity Holographic Society · p. 2
  • Holograms may integrate all five senses, with emotion-sensing wearables proposed to monitor mental health and facilitate social interaction.The paper presents multisensory communication as a way to enrich digital experiences beyond sight and hearing.Source: 1) High Fidelity Holographic Society · p. 2

2) Connectivity for All Things:

  • Future society may require an order-of-magnitude or greater increase in interconnectivity, spanning critical infrastructure and networks beyond terrestrial systems.Examples include terrestrial–satellite integration for global broadband and communications from UAVs; critical infrastructure also raises security and sensor-reliability needs.Source: 2) Connectivity for All Things · p. 2

3) Time Sensitive/Time Engineered Applications: … II. 6G USE CASES AND TECHNICAL REQUIREMENTS

The paper frames 6G as a response to applications needing fast, reliable information delivery, and takes a holistic top-down view of the technical and architectural changes required. It distinguishes practically realizable systems from theoretical possibilities.

3) Time Sensitive/Time Engineered Applications:

  • Time-critical, tactile services need guaranteed capacity and timely arrivals; current 5G cannot fully virtualize remote skills under the 1 ms human-reaction limit, motivating leaner 6G architectures and processing.The passage identifies tactile internet as supporting real-time touch and actuation, beyond conventional audio and video communication.Source: 3) Time Sensitive/Time Engineered… · p. 2

B. Literature Review: 6G Vision and Performance Aspects

  • Prior 6G surveys cover use cases, system requirements, KPIs, and enabling technologies, including holographic radio, THz communications, IRS, and OAM,, [15] [18].The paper notes that its literature review is necessarily incomplete and treats these studies as important examples; technology-specific work is also reviewed in related sections.Source: 3) Time Sensitive/Time Engineered… · pp. 2–3, B. Literature Review: 6G Vision and… · p. 3
  • Specific 6G studies address impaired waveforms at 90–200 GHz, molecular communication, wireless IoT powering [22], AI-based network management, collaborative AI [24], and THz RF hardware.Together with the current paper, these studies contribute to the 6G vision summarized in Fig. 1.Source: B. Literature Review: 6G Vision and… · p. 3

C. Contributions of the Paper

  • The paper takes a holistic, top-down view, linking 6G application capabilities and spectrum opportunities to deployment, PHY, hardware, propagation, and real-time processing challenges.It notes that bands at 100 GHz and above pose hardware and system-component challenges that limit where their spectrum can be used.Source: C. Contributions of the Paper · p. 3
  • 6G core and transport networks need flattening or major reduction because current architecture cannot meet sub-millisecond service latency, while best-effort TCP/IP cannot guarantee application QoS.The proposed direction includes virtualizing transport over existing fiber, isolating it with SDN and virtualization, and enabling core functions as microservices on demand.Source: C. Contributions of the Paper · p. 3
  • The paper assesses the strengths and weaknesses of proposed concepts and distinguishes practical realizability from theoretical possibility to identify research challenges and innovation opportunities.The authors state that, to their knowledge, a contribution of this holistic type is missing from the literature.Source: C. Contributions of the Paper · p. 3

D. Organization of the Paper

  • The paper proceeds from 6G vision, use cases, and KPIs to frequency bands, network architecture, PHY techniques, propagation, transceivers, real-time processing, and conclusions.It also provides a comprehensive bibliography for further reading.Source: D. Organization of the Paper · p. 3

II. 6G USE CASES AND TECHNICAL REQUIREMENTS

  • 6G use cases require instantaneous, extremely high-speed wireless connectivity, and the paper reviews published requirements and categorizes them by use case.It cites Network 2030 requirements published by ITU-T [28] alongside requirements from other cited sources.Source: §II · pp. 3–4

A. Use Case 1: Holographic Communications · B. Use Case 2: Tactile and Haptic Internet Applications

Holographic and tactile applications demand very high data rates, ultra-low latency, and coordinated delivery, while also imposing security, reliability, synchronization, and practical implementation challenges. Their requirements vary by application, from immersive displays and remote surgery to industrial automation and autonomous driving.

A. Use Case 1: Holographic Communications

  • Hologram data rates range from tens of Mbps to 4.3 Tbps, while interactive experiences require ultra-low latency and synchronized multisensory streams.Adding haptic capability can require sub-millisecond latency.Source: A. Use Case 1: Holographic… · p. 4
  • Holographic services must coordinate dependent co-flows and protect them against security failures, while resilience, computation, compression, bandwidth, and latency must be balanced.Remote surgery makes integrity and security vital; higher compression can reduce bandwidth needs but increase computation and latency.Source: A. Use Case 1: Holographic… · p. 4
  • Hologram realization remains constrained by computation-intensive generation, specialized recording setups, absent display-data standards, limited HMDs, and mobile GPU and battery limits.Mobile-device GPU performance is typically 1/40-th of an average personal computer GPU, and established JPEG/MPEG compression is unsuitable for holographic signal statistics.Source: A. Use Case 1: Holographic… · p. 5

B. Use Case 2: Tactile and Haptic Internet Applications

  • Remote industrial management needs real-time control, with advanced manufacturing robotics targeting 100 µs link latency and 1 ms round-trip reaction times.Robotics require guaranteed real-time control to avoid oscillatory movements; remote operators may use VR or holographic communications with tactile sensors and kinesthetic feedback.Source: B. Use Case 2: Tactile and Haptic… · p. 5
  • Autonomous-driving applications include driver assistance, platooning, and fully automated driving, with latency of a few ms likely needed for collision avoidance and remote driving.The passage notes that fully functional autonomous vehicles do not yet exist, so further requirements and applications may emerge.Source: B. Use Case 2: Tactile and Haptic… · p. 5
  • Tactile internet applications include remote robotic surgery, where surgeons use real-time audio-visual feeds and haptic information to operate remotely.The passage also identifies tele-diagnosis and telerehabilitation as potential healthcare applications.Source: B. Use Case 2: Tactile and Haptic… · p. 5
  • Tactile applications require about 1 ms end-to-end latency, potentially sub-ms feedback and synchronization, while application rates range from 1-5 Mbps for 1080p video to 0.5-2 Gbps for point-cloud holograms.The passage gives 15-25 Mbps for 4K 360° video and up to a few Tbps for large holograms; autonomous-driving sensors may aggregate 1 Gbps.Source: B. Use Case 2: Tactile and Haptic… · p. 5
  • Tactile services require security, criticality-based stream prioritization, and reliability beyond current wireless guarantees, reaching 99.999% or 99.99999% in cited cases.Industrial IoT may require 99.99999% reliability, while ultra-reliable transmissions are assumed to have a 99.999% success rate.Source: B. Use Case 2: Tactile and Haptic… · p. 6

C. Use Case 3: Network and Computing Convergence · D. Use Case 4: Extremely High Rate Information Showers · E. Use Case 5: Connectivity for Everything

The use cases envision coordinated edge computing, fibre-like information showers, and connectivity for extensive sensing and bio-related applications. Their requirements range from coordinated resource management to high rates, security, and potentially low latency.

C. Use Case 3: Network and Computing Convergence

  • Multiple edge sites can coordinate computation-intensive services such as AR/VR rendering, autonomous driving, and holographic communications, balancing workloads across edge and core clouds.A low-latency request may be routed to the nearest edge site, while demanding applications may involve multiple sites whose computing resources must be coordinated.Source: C. Use Case 3: Network and Computing… · p. 6
  • Edge-cloud coordination requires computing-aware facilities, joint network and compute scheduling, flexible resource-provider addressing, and fast routing or rerouting as loads change.Scheduling may be centralized or distributed, and any network node can become a resource provider.Source: C. Use Case 3: Network and Computing… · p. 6

D. Use Case 4: Extremely High Rate Information Showers

E. Use Case 5: Connectivity for Everything

  • Connectivity for everything includes real-time monitoring of buildings, cities, the environment, transport, and critical infrastructure, as well as wearable and implanted bio-sensors.Implanted sensors and intra-body communications extend connectivity needs beyond mMTC.Source: E. Use Case 5: Connectivity for… · p. 6

C. Use Case 3: Network and Computing Convergence

E. Use Case 5: Connectivity for Everything

  • These sensing applications require large aggregated data rates, strong security and privacy for medical data, and potentially low latency for urgent interventions.The paper notes that no systems or models yet exist to assess these data needs.Source: E. Use Case 5: Connectivity for… · p. 6

F. Use Case 6: Chip-to-Chip Communications … A. New Frequency Bands for 6G

The paper outlines 6G requirements and use cases spanning chip-to-chip and satellite-terrestrial communications, alongside the opportunities and constraints of higher-frequency spectrum. It identifies performance targets while distinguishing promising frequency bands from those limited by absorption, coexistence, or hardware challenges.

F. Use Case 6: Chip-to-Chip Communications

  • Chip-to-chip and inter-board wired links face bottlenecks above 100–1000 Gbps, motivating optical or THz wireless alternatives; nanonetwork KPIs include data rate, energy efficiency, reliability, and latency.Energy efficiency must account for any required receiver processing, while specific KPIs depend on chip implementations and applications.Source: F. Use Case 6: Chip-to-Chip… · p. 6

G. Use Case 7: Space-Terrestrial Integrated Networks

  • LEO satellite constellations integrated with terrestrial networks could provide global internet access, including on moving platforms, and satellite access without ground infrastructure.Potential benefits also include enriched internet paths across domains and ubiquitous edge caching and computing.Source: G. Use Case 7: Space-Terrestrial… · p. 7
  • Integrated satellite-terrestrial networks require flexible routing and addressing, capacity-aware inter-satellite and terrestrial links, satellite admission control, and onboard edge computing and storage; link distance sets a minimum delay.Onboard limitations challenge satellite edge computing and storage, while admission control requires satellites to know traffic loads across the space network.Source: G. Use Case 7: Space-Terrestrial… · p. 7
  • 6G requirements include peak data rate ≥1 Tbps (at least 50× 5G), user experience data rate at least 10× 5G, and application-dependent user-plane latency at least 40× better than 5G.Source: G. Use Case 7: Space-Terrestrial… · p. 7
  • 6G is expected to support mobility up to 1000 km/h and connection density potentially 10× that of 5G.Source: G. Use Case 7: Space-Terrestrial… · p. 7

A. New Frequency Bands for 6G

  • The 100 GHz–1 THz range is a 6G candidate, but high-absorption bands suit only short links, and existing services mean regulators may not provide all spectrum or allocate it contiguously.Passive services are especially interference-sensitive and require protections such as guard bands, emission and power limits, and terrestrial-beam restrictions.Source: A. New Frequency Bands for 6G · p. 7
  • The 140–350 GHz W1 sub-THz band offers tens of GHz of unused bandwidth and room for ultra-massive MIMO arrays; higher windows bring greater absorption, while optical links are becoming realistic integration options.Operation up to 10 THz is possible but faces formidable hardware challenges; optical options include laser diodes and LEDs for visible-light communications.Source: A. New Frequency Bands for 6G · p. 7

B. 6G Deployment Scenarios … 3) Wireless Personal Area Networks (WPANs):

6G deployment scenarios span existing and new frequency bands, from THz-driven hot spots to industrial networks and short-range WPAN/WLAN links. Their feasibility depends on coverage distance, reliability, latency, and link budget.

B. 6G Deployment Scenarios

  • 6G deployment scenarios include new THz-driven uses, IoT below 1 GHz, and autonomous-driving cellular V2X combining microwave and mmWave bands.Some 5G applications will continue in existing bands, which may be re-farmed to 6G over time.Source: B. 6G Deployment Scenarios · p. 8

1) Hot Spot Deployments:

  • MmWave and THz systems, including window W1, suit very-high-data-rate hot spots deployed indoors or outdoors.Ubiquitous deployment is uneconomical because outdoor coverage radius is about 100 m and indoor coverage is shorter, due to free-space pathloss and molecular absorption. Aggregating windows can increase bandwidth while shortening feasible distance; realistic planning must also account for obstructions, scattering, and other effects beyond the free-space-only link budgets.Source: 1) Hot Spot Deployments · p. 8

2) Industrial Networks:

  • Industrial networks are envisaged as private deployments prioritizing extreme reliability and ultra-low latency for sensor-robot, robot-coordination, and operator-robot communications.Most current commercial deployments targeting ultra-high reliability use 3.4–3.8 GHz, where diffraction efficiency makes propagation relatively rich.Source: 2) Industrial Networks · p. 8
  • Over the next decade, industrial networks are expected to combine mmWave with bands below 6 GHz to support high data rates alongside real-time control and AI.These capabilities would enable transmission and processing of high-definition visual data for digital twins of machines and operations, as well as remote troubleshooting; preliminary studies show possibilities and challenges for mmWave integration.Source: 2) Industrial Networks · p. 8

3) Wireless Personal Area Networks (WPANs):

  • WPAN and WLAN links can connect laptops and access points, kiosks and receivers, AR/VR wearables and modems, or proposed “infostations”.The source’s examples include links between a laptop and an access point, and between an information kiosk and a receiver.Source: 3) Wireless Personal Area Networks… · pp. 8–9
  • WPAN links may be under 0.5–1 m and WLAN links up to 30 m; any THz window may suit them if the link budget covers path loss and implementation technologies exist.Source: 3) Wireless Personal Area Networks… · p. 9

4) Autonomous Vehicles and Smart Railway Networks: … 4) 360-Cybersecurity & Privacy-by-Engineering-Design:

6G is envisioned as building on 5G’s flexible, service-driven approaches while introducing new network design principles, protocols, and security and privacy mechanisms. The proposed changes address both emerging applications and the evolution of network architecture.

4) Autonomous Vehicles and Smart Railway Networks:

  • THz links may support high-rate railway safety and passenger-data transmission, but train mobility makes them sensitive to beamforming errors and Doppler spread.Prediction can guide beam steering because train speed is nearly constant, but narrow beams and high beamforming gain make small prediction deviations consequential. High-frequency cabin access can also connect user equipment to antennas aggregating passenger data.Source: 4) Autonomous Vehicles and Smart… · p. 9
  • 6G could enable information sharing between autonomous vehicles and infrastructure, though complicated traffic and short ranges raise doubts about THz suitability.Source: 4) Autonomous Vehicles and Smart… · p. 9

IV. 6G RADIO AND CORE NETWORK ARCHITECTURES: DESIGN PRINCIPLES AND FUNDAMENTAL CHANGES

  • 6G is framed as consolidating 5G’s flexible topologies, network slicing, control/user plane separation, and MEC while introducing new transport-architecture and design approaches.The service-driven architecture’s software components and APIs support an open innovation community and faster deployment; the discussion turns to radio and core network architecture changes for emerging use cases and requirements.Source: §IV · p. 9, 4) Autonomous Vehicles and Smart… · p. 9

A. 6G Network Design Principles

  • 6G network design calls for convergence of non-3GPP wired and radio systems, mutual or 3GPP-driven security, and non-IP protocols suited to wireless-edge traffic.Convergence can balance traffic across networks with different loads and hedge delivery across technology families for resilience; more than 50% of networking traffic originates in or terminates at the wireless edge.Source: A. 6G Network Design Principles · p. 9

3) Information-Centric & Intent-Based Networks (ICN):

  • ICN separates content from its location identifier, while intent-based networking uses policies, continuous monitoring, and real-time optimization to keep network and service behavior aligned with intent.ICN names content using an abstract naming convention rather than IP addressing; intent-based management configures end-to-end softwarized infrastructure and adapts to changes in network or service state.Source: 3) Information-Centric & Intent-Based… · pp. 9–10

4) 360-Cybersecurity & Privacy-by-Engineering-Design:

  • 6G’s 360-cybersecurity approach extends protection from architecture and protocols to embedded software, while privacy-by-engineering-design builds privacy checks into protocols and architecture.For example, a camera would stream footage only when networking-level and possibly contextual privacy requirements are met.Source: 4) 360-Cybersecurity &… · p. 10
  • 6G should provide mechanisms to integrate emerging and not-yet-invented technologies more quickly, including quantum, distributed ledger technologies, and AI.Quantum is proposed for tamperproof infrastructure, cryptographic key exchange, and faster optimization; distributed ledgers can provide immutable provenance and support trust across telecom partners, while distributed AI and explainable AI address atomized networks and regulatory requirements.Source: 4) 360-Cybersecurity &… · p. 10

B. Opportunities for Fundamental Change … 4) AI-Native Design Enabling Human-Machine Teaming:

The paper identifies major infrastructure changes and native AI integration as opportunities for 6G, alongside network designs intended to support societal needs and human-machine teaming.

B. Opportunities for Fundamental Change

1) Removal/Reduction of the Transport Network:

  • Operators could reduce or remove private transport networks and carry cellular traffic over sliced Internet fiber, focusing on the wireless edge and control plane.The approach requires policy and operational changes, though the supporting technologies are available.Source: 1) Removal/Reduction of the Transport… · p. 10

2) Flattened Compute-Storage-Transport:

  • 6G could flatten compute, storage, and transport through virtualized fiber, microservice-based core functions, and separate local GPU and nearby MEC CPU execution.The paper also identifies open source as a growing 6G component for security, faster innovation, and contributions across the communications and computer science communities. Open data and automated access to operational data are needed for algorithm training while protecting network security and customer privacy.Source: 2) Flattened Compute-Storage-Transport · pp. 10–11

4) AI-Native Design Enabling Human-Machine Teaming:

  • 6G must natively fit distributed AI, transfer learning, and ensemble techniques into telecom architectures while making consumer-facing AI decisions explainable and policy-compliant.Explainable AI can help people understand AI decisions and trust them; traditional deep-learning systems may operate as black boxes.Source: 4) AI-Native Design Enabling… · p. 11
  • Future networks could use AI to turn telecom innovations into validated code and deploy it across infrastructure, enabling human-machine teaming.The paper notes that this AI-designed-network vision may not arrive with 6G.Source: 4) AI-Native Design Enabling… · p. 11

5) Human-Centric Networks:

  • 6G could move beyond static cell- and device-centric designs toward societally aware, adaptable networks that address differing user requirements and societal disruptions.The paper cites COVID-19 resource shifts, privacy concerns, and security breaches as shortcomings, and calls for analysis and communication of effects on health and well-being.Source: 4) AI-Native Design Enabling… · pp. 11–12

4) AI-Native Design Enabling Human-Machine Teaming:

  • 3–6 ms of added latency can arise in non-congested multi-operator 5G backhaul, limiting ultra-low-latency quality-of-service assurances for endpoints on different operators.A proposed local-breakout network could reduce or remove transport fiber; sliced ISP infrastructure and end-to-end orchestration, potentially using a distributed ledger, are also needed.Source: 4) AI-Native Design Enabling… · p. 12

V. NEW PHYSICAL LAYER TECHNIQUES FOR 6G · A. Modulation, Waveforms, and Codes · 1) Multicarrier Techniques:

The section surveys modulation, waveform, and coding directions for 6G, then focuses on multicarrier techniques and the trade-offs and research challenges that shape their use. It places these topics within a broader physical-layer discussion spanning antennas, multiple access, optical communications, AI/ML, and vehicular communications.

V. NEW PHYSICAL LAYER TECHNIQUES FOR 6G

  • The broader physical-layer discussion covers modulation, waveforms, coding, multiple-antenna and multiple-access techniques, free-space optical communications, AI/ML applications, and vehicular communications; it omits several topics for space reasons.The section also discusses ultra-massive MIMO, distributed antenna systems, intelligent-surface-assisted communications, and orbital angular momentum systems.Source: §V · p. 12

1) Multicarrier Techniques:

  • 6G’s more heterogeneous applications may require a wider range of modulation formats than 5G, although OFDM may remain for backward compatibility despite its non-ideal-channel drawbacks.OFDM’s popularity stems from its capacity optimality over frequency-selective channels and its use in 4G and 5G.Source: 1) Multicarrier Techniques · p. 12
  • At mmWave and THz, OFDM’s frequency-dispersion sensitivity, cyclic-prefix spectral-efficiency loss, and high PAPR become more critical, with Doppler, phase noise, and inter-user subcarrier interference further reducing performance.High PAPR requires linear power amplifiers operating with backoff, which sacrifices efficiency, and high-resolution converters whose energy use and cost rise with bandwidth; modulation must balance capacity, converter resolution, and equalization complexity.Source: 1) Multicarrier Techniques · pp. 12–13
  • Multicarrier designs trade off critical sampling, complex-domain orthogonality, and time-frequency localization, which cannot all be achieved simultaneously under the Balian-Low theorem.Options include orthogonal methods such as FMT, UFMC, and FBMC; bi-orthogonal methods such as windowed OFDM; and non-orthogonal schemes such as GFDM and faster-than-Nyquist signalling.Source: 1) Multicarrier Techniques · p. 13
  • OTFS maps QAM to the delay-Doppler rather than time-frequency domain, using frequency dispersion as diversity and enabling flexible multiplexing across users with different delay profiles and Doppler spectra.Real-time prototypes exist, but efficient equalization—particularly for multiple-antenna systems—and implementation remain research needs.Source: 1) Multicarrier Techniques · pp. 13–14
  • Massive-IoT and internet-of-bio-things devices need modulation that minimizes transmitter and receiver energy, since energy-harvesting or long-life nodes cannot practically use flash signalling’s high PAPR and precise synchronization.Clock-free receivers, molecular-communication ideas, and passive backscatter may help reduce energy use; different modulation methods may be used for uplink and downlink.Source: 1) Multicarrier Techniques · p. 14
  • Massive antenna systems make pilot overhead for channel-state information acquisition potentially prohibitive, motivating adaptive pilot spacing, exploitation of limited angular spread, contamination reduction, and further work on feedback and machine-learning methods.Reciprocity calibration and spatial or temporal extrapolation show promise, but system-level imperfections limit their gains.Source: 1) Multicarrier Techniques · p. 14

2) Advances in Coding: … 2) Intelligent Surface-Assisted Communications:

6G designs must balance reliability, latency, energy, and implementation complexity across coding and antenna systems, while intelligent surfaces introduce new control, deployment, and transport-network challenges.

2) Advances in Coding:

  • 6G short-packet coding must balance block length, error robustness, latency, and energy, combining forward error correction with iterative retransmission or feedback and ML-based methods.Short-block LDPC and Polar codes are used in 5G uplink/downlink channels; short blocks risk errors and ARQ delays, while longer blocks add latency, and ARQ can increase energy use in low-energy devices –.Source: 2) Advances in Coding · pp. 14–15

1) Ultra Massive MIMO Systems:

  • 6G antenna arrays may scale by another order of magnitude, requiring beamforming choices that balance gain against the complexity, energy use, and cost of digital processing at mmWave frequencies.Digital beamforming is popular in the 3.4–3.8 GHz C-band, while commercial 24.5–29.5 GHz deployments mostly use analog beamforming; digital beamforming for 64 antennas at 28 GHz has also been demonstrated. Hybrid beamforming balances analog and digital processing.Source: 1) Ultra Massive MIMO Systems · p. 15
  • Electrically ultra-massive arrays amplify near-field wavefront curvature, differential shadowing, and beam squint, demanding suitable architectures and real-time signal-processing algorithms.Spatial modulation is a lower-complexity alternative that conveys part of the coded symbol through the antenna index; related work explores channel estimation, differential and hybrid methods, and visible-light communication [100]–.Source: 1) Ultra Massive MIMO Systems · p. 15
  • Cell-free massive MIMO could extend distributed antenna systems below 6 GHz, but retaining theoretical spectral- and energy-efficiency gains in realistic mobile scenarios and synchronizing distributed nodes remain challenges.The scenarios may span hundreds of meters and involve UE and scatterer mobility; real-time systems must synchronize multiple access points, UEs, and the central processing unit.Source: 1) Ultra Massive MIMO Systems · p. 15

2) Intelligent Surface-Assisted Communications:

  • Large intelligent surfaces can act as ultra-massive arrays, while intelligent reflecting surfaces quasi-passively redirect signals using tunable phase shifters; steering, interference, and energy optimization need further study.The literature also uses names such as reconfigurable intelligent surfaces and holographic beamforming [113] [114] [115].Source: 2) Intelligent Surface-Assisted… · pp. 15–16
  • Surface deployment requires evaluating coverage and cost against active relays and passive reflectors, and assessing reliability under element failures and environmental conditions that can cause misalignment and maintenance overhead.Relevant conditions include cracks, temperature variation, rain, and wind; deployment analysis should account for circuit performance, electromagnetic behavior, space rental, and ongoing maintenance.Source: 2) Intelligent Surface-Assisted… · p. 16
  • IRS control requires efficient BS–surface and UE–surface signaling, adaptation to traffic changes, recalibration or calibration-free operation, and a dedicated transport link whose bandwidth and control-plane latency are design challenges.The required link throughput depends on surface element count, control bits per element, refresh rates, frame timing, and the number of UEs; added links increase control-plane, but not user-plane, latency.Source: 2) Intelligent Surface-Assisted… · p. 16

3) OAM-Based Systems: · C. Multiple Access Techniques

The section surveys OAM’s potential and practical limits, alongside the need for new multiple-access approaches that scale to massive low-duty-cycle connectivity and improve high-spectral-efficiency performance in 6G.

3) OAM-Based Systems:

  • OAM multiplexes beams by imposing phase twists that make different modes orthogonal, and the modes can be separated using analog means such as spiral phase plates.Source: 3) OAM-Based Systems · p. 16
  • OAM suits line-of-sight settings such as data centers and wireless backhaul, but its near-field multiplexing principle limits range; electrically larger antennas make it better suited to high mmWave, THz, and free-space optics.Source: 3) OAM-Based Systems · pp. 16–17
  • Practical OAM use requires greater robustness to multipath, orientation misalignment, and other impairments; preliminary work addresses multipath and turbulence, but further work is needed.Source: 3) OAM-Based Systems · p. 16

C. Multiple Access Techniques

  • Rate splitting (RS) divides UE messages into common and private parts, decodes common streams, applies successive interference cancellation, then decodes private streams, flexibly decoding some interference and treating some as noise.The common and intended private message parts are combined to reconstruct each receiver’s original message.Source: C. Multiple Access Techniques · p. 17
  • Simplified NOMA or RS may contend for 6G; NOMA was intended for 5G but omitted from early releases amid the rush to finish specifications.Source: C. Multiple Access Techniques · p. 17
  • 6G research should improve high-spectral-efficiency access toward theoretical limits while accounting for practical precoding constraints and available CSI.Source: C. Multiple Access Techniques · p. 17

D. Free-Space Optical Communications … A. MmWave and THz Propagation Channels

The sections describe optical, AI/ML, and vehicular approaches to 6G, then examine propagation across sub-6 GHz, mmWave, and THz bands. They identify substantial challenges in channel behavior, modeling, measurement, and system integration.

D. Free-Space Optical Communications

  • Optical links promise extremely high data rates over small-to-medium distances when LOS is guaranteed, but mobility, blockage, and cellular/WiFi integration remain challenges.LED-based VLC/LiFi can reuse lighting for downlink while traditional radio provides uplink; laser-based systems offer higher rates but narrow beams suit fixed wireless scenarios, and NLOS operation may require different rates and signal processing.Source: D. Free-Space Optical Communications · p. 17

E. Applications of AI and ML

  • ML may support PHY detection, decoding, precoding, and beam or antenna selection, and channel estimation and prediction in systems with short codewords, low-resolution hardware, and large, time-varying arrays.Research questions include algorithm choice, training-data needs, transferability across environments, and explainability.Source: E. Applications of AI and ML · p. 17

F. Vehicular Communications

  • Vehicle sensors can produce raw aggregate data rates up to 1 Gbps, motivating below-6-GHz links for reliability and mmWave links for Gbps rates.Vehicles may use sensor sharing and fusion for automated driving, especially in congested scenarios.Source: F. Vehicular Communications · pp. 17–18
  • V2X research must address propagation and channel non-stationarity, MEC integration for latency, sensor-fusion processing trade-offs, and faster beam management for mobile, blockage-prone vehicles.Old cars may limit V2X/V2I gains until late 2030’s; the described use case combines mmWave links for platoon sensing and controller links with sub-6 GHz broadcasts for basic information.Source: F. Vehicular Communications · p. 18, Figure 8 · p. 18

VI. PROPAGATION CHARACTERISTICS OF 6G SYSTEMS

  • The propagation overview covers sub-6 GHz, mmWave, and THz channels, including ultra-massive MIMO, distributed antenna, V2V/V2I, industrial, UAV, and wearable channels.Full-duplex and device-to-device channels are omitted due to space constraints.Source: §VI · p. 18

A. MmWave and THz Propagation Channels

  • At mmWave and THz, propagation is shaped by frequency-dependent path loss and absorption, weaker diffraction, scattering and penetration effects, and frequency-scaled Doppler shifts.Atmospheric attenuation is approximately 10 dB/km at the 60 GHz oxygen line and can exceed 100 dB/km at multiple peaks from 100–1000 GHz; band selection must align with link distance.Source: A. MmWave and THz Propagation Channels · pp. 18–19
  • THz channel models remain limited because measurements and validation are sparse, while suitable high-resolution sounders are difficult, expensive, and time-consuming to build.Existing models are often scenario-specific; broader environments require mixed deterministic-stochastic modeling and extensive measurements, and high-resolution terrain databases are also lacking.Source: A. MmWave and THz Propagation Channels · p. 19

B. Propagation Channels for Distributed Antenna Systems · C. Ultra Massive MIMO Propagation Channels

Distributed-system channel models must capture correlated links across users and base stations, while ultra-massive arrays require propagation models that account for near-field effects and spatially varying channel behavior.

B. Propagation Channels for Distributed Antenna Systems

  • Distributed base stations may use enhanced cloud RAN, CoMP, or cell-free massive MIMO; most deployments are expected below 6 GHz, with mmWave considered for high data rates alongside reliability.The paper notes that investigations of mmWave distributed systems remain limited.Source: B. Propagation Channels for… · p. 19
  • Correlations among multiple UE links can persist across distant base stations, and common clusters can model shared propagation effects.Correlated angular spreads, delay spreads, and mean directions have been measured; shadowed common clusters can jointly affect received power and multiple UEs’ angular and temporal dispersion, as used in COST 2100.Source: B. Propagation Channels for… · pp. 19–20

C. Ultra Massive MIMO Propagation Channels

  • Ultra-massive MIMO is envisioned at high mmWave and THz frequencies with potentially thousands of elements, while sub-6 GHz arrays offer research opportunities but are challenging to deploy [149].A taxonomy of THz operation using arrays of subarrays is provided in.Source: C. Ultra Massive MIMO Propagation… · p. 20
  • Large apertures can produce spatial non-stationarities and wavefront-curvature effects that weaken channel hardening and favorable propagation in ultra-massive MIMO.Different array regions may see partially unique scatterers and users; wavefront curvature can vary both multipath phases and amplitudes across the array.Source: C. Ultra Massive MIMO Propagation… · p. 20
  • As arrays grow below 6 GHz, the plane-wave assumption can break down when users and scatterers lie within the Fresnel zone, whose boundary is the Fraunhofer distance df = 2D^2/λ.Increasing array dimension D at fixed wavelength λ increases the likelihood of near-field propagation.Source: C. Ultra Massive MIMO Propagation… · p. 20
  • Near-field channel impulse responses require models grounded in near-field propagation physics, with measurement-based studies documenting spatial non-stationarity and geometrical modeling approaches [149].The COST 2100 extension models dynamic cluster appearance and disappearance through separable scatterer visibility regions unique to both link ends [149].Source: C. Ultra Massive MIMO Propagation… · p. 20
  • Characterizing very large-array channels is constrained by measurement runs lasting tens of minutes or longer, while fully parallel measurements are limited by cost and energy use.Long runs assume quasi-static channels and restrict the range of feasible measurement scenarios.Source: C. Ultra Massive MIMO Propagation… · p. 20

D. Propagation in Industrial Environments … G. Wearable Propagation Channels

Propagation differs substantially across industrial, aerial, vehicular, and wearable settings, with mobility, blockage, environment, and device placement shaping channel behavior. The passages report established findings alongside gaps in dynamic measurements and generalizable channel models.

D. Propagation in Industrial Environments

  • Industrial measurements show shadow-fading decorrelation distances of about 15 m and 30 m at 2.37 and 5.4 GHz, exceeding 3GPP’s 6 m and 10 m values [163],.At 3.7 and 28 GHz, delay spread showed no substantial difference, while further work is needed on time-varying industrial channels below 6 GHz and at mmWave frequencies.Source: D. Propagation in Industrial… · p. 21

E. UAV Propagation Channels

  • Air-to-air UAV channels typically approximate free space, whereas air-to-ground channels have more scattering, two-path fluctuations, shadowing, Ricean fading, and high Doppler.Airframe obstruction can add shadowing, and mmWave empirical studies remain scarce; calibrated ray tracing can predict 60 GHz UAV channels.Source: E. UAV Propagation Channels · p. 21
  • Stratospheric balloons at about 20 km yield roughly 0.1 ms two-way propagation delay, versus 2–8 ms for LEO and 240 ms for geostationary satellites, supporting tight-latency services.The cited commercial 4G service was launched by Loon and Telkom in Kenya in June 2020.Source: E. UAV Propagation Channels · p. 21

F. Vehicular Propagation Channels

  • Below 6 GHz, vehicular channels are nonstationary, can experience up to 4× conventional cellular Doppler spread, and range from sparse highway to richer urban multipath.Multipath components have limited lifetimes, vehicle blockage affects path loss, and antenna position and pattern alter channel statistics; an obstructing truck caused 12–13 dB median loss,,.Source: F. Vehicular Propagation Channels · pp. 21–22
  • At 28 GHz, highway V2I measurements at 100 km/h found Doppler spread up to 10 kHz and estimated coherence time as low as 100 µs.MmWave vehicular evidence remains limited and often relies on stationary measurements; reported LOS blockage losses are 5–30 dB, while measured antenna excess path loss ranges from 10 to 25 dB depending on angle,. Delay-Doppler sparsity at 60 GHz may suit OTFS modulation.Source: F. Vehicular Propagation Channels · p. 22

G. Wearable Propagation Channels

  • Wearable channel models are not standardized, and existing fading models depend on measured body locations, antenna types, and frequency bands, limiting generalization.Measurements span narrowband 300 kHz–1 MHz at sub-1 and 2 GHz and ultra-wideband 499 MHz in C-band and 6–10 GHz; one study included 60 subjects.Source: G. Wearable Propagation Channels · p. 22

VII. REAL-TIME PROCESSING AND RF TRANSCEIVER DESIGN: CHALLENGES, POSSIBILITIES, AND SOLUTIONS · A. Implications of Increasing Carrier Bandwidths · B. Processing Aspects for mmWave and THz Frequency Bands

The section weighs the promise of wideband mmWave and THz systems against carrier aggregation, propagation, beamforming, and real-time processing constraints. It outlines candidate architectures and illustrates theoretically achievable capacity while emphasizing unresolved practical conditions.

A. Implications of Increasing Carrier Bandwidths

  • Aggregating carriers can provide wide bandwidths, but 100 carriers for 10 GHz would require calibration across them amid phase noise and band-dependent RF performance.A 1 GHz receiver bandwidth raises the noise floor by 10 dB relative to a 100 MHz carrier, degrading SNR; carrier count and operable bandwidth also depend on antenna-integrated RF circuits and EIRP safety limits.Source: A. Implications of Increasing Carrier… · p. 22

B. Processing Aspects for mmWave and THz Frequency Bands

  • THz free-space loss exceeds 100 dB at 10 m, constraining cell range; ultra-massive MIMO could help close the link budget, subject to energy and beamforming limits.These constraints make some services beyond the 140–350 GHz W1 window difficult to envisage, despite the need for Tbps connectivity.Source: B. Processing Aspects for mmWave and… · p. 23
  • Fully digital beamforming at THz is formidable, so analog or hybrid phased arrays appear likeliest short-to-medium-term options, though their processing architectures need redesign.Graphene plasmonic antennas, metamaterial antennas and hypersurfaces, and metasurface lenses are discussed; lenses could replace complex, power-hungry RF splitting, phase-shifting, and combining circuits with a passive device.Source: B. Processing Aspects for mmWave and… · p. 23
  • Real-time control of array RF interconnects and beamforming networks is harder at mmWave and THz because coherence times shorten, phase noise rises, and antenna counts increase.Fully connected hybrid architectures are cost prohibitive; the authors propose array-of-sub-arrays designs, baseband control of data streams and user interference, and low-resolution ADCs and DACs.Source: B. Processing Aspects for mmWave and… · p. 23
  • A peak capacity of 1 Tbps is theoretically achievable with 100 GHz bandwidth at 10 dB SNR under ideal transceiver and CSI assumptions.The toy example uses 4096 BS antennas, 16 UE antennas, and a 15 m link; reducing SNR to 3 dB causes an almost constant capacity loss across CDF values. The comparison at the same SNR levels with 1 GHz bandwidth yields less than a 100× capacity difference.Source: B. Processing Aspects for mmWave and… · p. 23, Figure 9 · p. 24
  • Realizing the theoretical capacity remains an open challenge: practical beamforming, sustained forward-link SNR near 10 dB, and modulation and coding gains must be established.The authors identify these as research questions for utilizing the large bandwidths available at THz frequencies.Source: B. Processing Aspects for mmWave and… · p. 24
  • Increasing THz capacity requires balancing bandwidth, MIMO layers, and SINR against declining power density, rising noise, uncertain channel rank, and network costs.More layers require ultra-massive arrays at both ends and sufficient channel rank; densification reduces competing users but increases operating expenditure and base-station coordination overhead.Source: B. Processing Aspects for mmWave and… · p. 24

C. RF Transceiver Challenges and Possibilities · D. Comments on Energy Consumption and Efficiency · VIII. CONCLUSIONS

The paper examines the hardware and energy-efficiency challenges of 6G, especially at THz frequencies, and outlines research directions across the system stack while distinguishing practical possibilities from theoretical ones.

C. RF Transceiver Challenges and Possibilities

  • Sub-6 GHz and mmWave base stations use radio-over-fiber, active integrated antennas, real-time conversion control, and cascaded amplifier stages, but these architectures do not transfer readily to THz.The illustrated architecture includes filtering and control circuits as well as TX/RX mixing and de-mixing operations.Source: C. RF Transceiver Challenges and… · p. 24, Figure 10 · p. 25
  • THz operation degrades receiver noise figure and transmitter output power, while reduced antenna aperture and wide bandwidth constrain link distances without coherent beamforming from thousands of elements.The text estimates that THz base stations may require thousands to tens of thousands of antenna elements.Source: C. RF Transceiver Challenges and… · p. 24
  • At 500 GHz, a 10,000-element array could fit in 3 cm × 3 cm, but 100 mW per transceiver implies 1 kW total and difficult heat dissipation.Sparse arrays with more than half-wavelength spacing may ease heat dissipation, but their side lobes could constrain spectrum sharing.Source: C. RF Transceiver Challenges and… · p. 24
  • SiGe technology is predicted to reach fmax near 2 THz at 5 nm and support integrated amplifiers and oscillators up to about 1 THz; 25 nm InP HEMTs have demonstrated 800 GHz amplifiers and mixers.The paper presents InP for the highest-frequency parts as a near-term option and anticipates that 5 nm SiGe could increase integration and reduce costs.Source: C. RF Transceiver Challenges and… · pp. 24–25
  • Distributed PLLs avoid power-hungry central 500 GHz LO distribution and enable LO beamforming, but PLL count trades off notch depth and simultaneous-beam performance [191]; THz tuning needs alternatives to varactors.The text notes that many PLLs can increase total oscillator energy and favor low-power reference distribution, while varactor quality factor falls with frequency.Source: C. RF Transceiver Challenges and… · p. 25

D. Comments on Energy Consumption and Efficiency

  • Large arrays and smaller cells can improve energy efficiency through directed transmissions, but worsening PA efficiency and UE noise figure at higher frequencies offset some gains.Despite high array power consumption, the paper expects energy per bit to drop by orders of magnitude as data rates rise; processing the data may become the bottleneck.Source: D. Comments on Energy Consumption and… · p. 25
  • Technology scaling alone may yield only 4× energy reduction in the coming decade, as baseband improvements slow, voltage scaling nears transistor thresholds, and parallelism offers less benefit.The text reports baseband improvements slowing to 10× per decade, versus 12× per decade for GPUs, and says UE active-core counts have remained essentially constant for two generations.Source: D. Comments on Energy Consumption and… · pp. 25–26

VIII. CONCLUSIONS

  • The paper takes a holistic top-down approach across 6G use cases, requirements, deployment, applications, core networks, and PHY, distinguishing theoretical possibilities from practical achievability.It emphasizes backward compatibility because devices will fall back to 5G or 4G with coverage, while 6G RAN, core, waveform, and coding choices face compatibility trade-offs.Source: §VIII · p. 26

Related papers

Loading 2008.03213v2…