PaperlayerView exact revision on arXiv

Source-linked AI summary

What Will the Future of UAV Cellular Communications Be? A Flight from 5G to 6G

Giovanni Geraci, Adrian Garcia-Rodriguez, M. Mahdi Azari, Angel Lozano, Marco Mezzavilla, Symeon Chatzinotas, Yun Chen, Sundeep Rangan, Marco Di Renzo

arXiv:2105.04842v1cs.ITcs.NIeess.SP

TL;DR

The paper examines how cellular networks can satisfy UAVs’ stringent connectivity, control, and capacity demands from 5G through 6G. It combines realistic case studies and original results to assess current NR capabilities and prospective 6G enablers, finding encouraging aerial coverage alongside substantial technological hurdles.

  • Problem

    Network-connected UAVs require reliable connectivity, control links, and high data-transfer capacity despite blocked air-to-ground links and coverage gaps.

  • Method

    The paper surveys industrial and 3GPP developments, evaluates 5G NR technologies through realistic case studies, and examines 6G enablers for UAV communications.

  • Results

    At least 98% of UAVs at 30 m achieve SNR > −5 dB with standard NR mmWave base stations, while coverage improves substantially at 120 m.

  • Takeaways & Limitations

    5G NR can support aerial connectivity, while 6G technologies such as NTNs, cell-free architectures, AI, RISs, and THz communications could improve coverage, reliability, and capacity.

  • Takeaways & Limitations

    THz UAV communications face channel-characterization, power-consumption, and size-and-weight constraints, including costly high-bandwidth analog-to-digital converters.

Abstract

from arXiv · show

What will the future of UAV cellular communications be? In this tutorial article, we address such a compelling yet difficult question by embarking on a journey from 5G to 6G and sharing a large number of realistic case studies supported by original results. We start by overviewing the status quo on UAV communications from an industrial standpoint, providing fresh updates from the 3GPP and detailing new 5G NR features in support of aerial devices. We then show the potential and the limitations of such features. In particular, we demonstrate how sub-6 GHz massive MIMO can successfully tackle cell selection and interference challenges, we showcase encouraging mmWave coverage evaluations in both urban and suburban/rural settings, and we examine the peculiarities of direct device-to-device communications in the sky. Moving on, we sneak a peek at next-generation UAV communications, listing some of the use cases envisioned for the 2030s. We identify the most promising 6G enablers for UAV communication, those expected to take the performance and reliability to the next level. For each of these disruptive new paradigms (non-terrestrial networks, cell-free architectures, artificial intelligence, reconfigurable intelligent surfaces, and THz communications), we gauge the prospective benefits for UAVs and discuss the main technological hurdles that stand in the way. All along, we distil our numerous findings into essential takeaways, and we identify key open problems worthy of further study.

I. INTRODUCTION

UAVs are moving toward widespread industrial and societal use, but ground-oriented cellular networks face severe aerial connectivity challenges. This tutorial combines industrial updates, original research results, and a 5G-to-6G outlook to assess solutions and open problems.

  • UAVs are expected to support applications including monitoring, emergency assistance, high-resolution mapping, inspection, and aerial radio access nodes.
  • Reliable connectivity must support latency-sensitive command-and-control links and increasingly large real-time payload-data transfers.
  • Downtilted ground BSs expose UAVs to sidelobes, fluctuating signals, widespread line-of-sight interference, and frequent or unnecessary handovers.
  • The article merges industrial and academic perspectives, using 3GPP updates and original results to evaluate 5G NR features and forecast UAV communications beyond 5G.
  • Its scope covers 5G NR status and features, mMIMO, mmWave, UAV-to-UAV links, future use cases, five 6G paradigms, and open research questions.

A. 3D Aerial Highways and the Delivery of Tomorrow

Future UAV applications include autonomous delivery, aerial traffic management, and demanding video-based services. These uses require reliable, low-latency connectivity for many UAVs operating in shared airspace.

  • A. 3D Aerial Highways and the Delivery of Tomorrow: Cargo UAVs are envisioned to deliver packages through regulated 3D aerial highways, with sense-and-avoid systems and coordinated traffic management.
  • A. 3D Aerial Highways and the Delivery of Tomorrow: Safe cargo-UAV operations require continuous data collection on speed, location, battery levels, and environmental conditions to trigger control commands.
  • A. 3D Aerial Highways and the Delivery of Tomorrow: Future networks may need end-to-end latencies of tens of milliseconds and minimal connectivity gaps while supporting large numbers of connected UAVs.
  • A. 3D Aerial Highways and the Delivery of Tomorrow: 5G UAV use cases broadly comprise command-and-control and payload-data links, with requirements spanning bit rate, latency, reliability, speed, altitude, and positioning accuracy.
  • A. 3D Aerial Highways and the Delivery of Tomorrow: Remote steering may require uplink video feedback, with specifications depending on whether the controller operates within or beyond visual line of sight.
  • A. 3D Aerial Highways and the Delivery of Tomorrow: Panoramic 8K live video from UAVs to remote virtual-reality users combines high-rate, low-latency, and accurate-positioning requirements.

3) UAVs as Radio Access Nodes:

UAVs acting as radio access nodes can be rapidly deployed for disaster, surveillance, emergency, and hotspot coverage, but aerial operation also creates access, interference, and mobility challenges. NR features such as beamformed SSBs, mMIMO, and dual connectivity are presented as remedies.

  • 3) UAVs as Radio Access Nodes:: UAVs carrying BSs, denoted UxNBs, can provide rapidly deployable coverage for disasters, border surveillance, emergency assistance, and hotspot events.
  • 3) UAVs as Radio Access Nodes:: NR combines sub-6 GHz mMIMO and transmissions above 24.25 GHz through mmWave to support aerial connectivity requirements.
  • 3) UAVs as Radio Access Nodes:: LTE/NR dual connectivity improves C2 reliability through redundant links and can offload UAV uplink traffic from LTE resources serving ground users.
  • 3) UAVs as Radio Access Nodes:: UAVs can avoid typical ground-user uplink power limitations because they generally transmit at much lower power.
  • A. Tackling the Initial Access and Cell Selection Challenges: NR beam-swept SSBs can direct synchronization signals toward the sky, enabling smoother RSRP transitions and easier initial cell selection and handovers.
  • A. Tackling the Initial Access and Cell Selection Challenges: In a representative UMa network, UAVs associate with non-nearest cells and can experience many handovers as adjacent positions change the strongest serving cell.

B. Tackling the Interference Challenge

NR massive MIMO improves UAV service quality by increasing directionality, suppressing intercell interference, and mitigating the effects of pilot contamination, while UAVs can still degrade GUE performance.

  • B. Tackling the Interference Challenge: NR mMIMO is essential for reliably serving worst-performing UAVs in fully loaded networks through beamforming and spatial multiplexing.The evaluated mMIMO system uses 128 RF chains, an 8×8 planar array, and zero-forcing precoding with perfect CSI.
  • B. Tackling the Interference Challenge: mMIMO null steering places up to 16 radiation nulls toward the most interfering UAVs using statistical CSI.The nulls operate during both uplink SRS reception and data transmission to suppress harmful interference.
  • B. Tackling the Interference Challenge: 42% to 85%: in UMi networks, the share of UAVs reaching 100 kbps doubles with NR mMIMO.Only 33% reach this rate in UMi single-user networks, while null steering nearly serves all UAVs in fully loaded UMa networks.
  • B. Tackling the Interference Challenge: UAV-aware intercell interference suppression can nearly double UL rates for coexisting aerial and ground devices.Higher network density worsens UAV performance because more base stations exchange interference with neighboring line-of-sight UAVs.
  • B. Tackling the Interference Challenge: UAVs affect GUE performance in both directions: uplink interference harms reception, while uplink SRS pilot contamination affects downlink transmission.High-altitude UAV SRS transmissions interfere with multiple base stations receiving GUE reference signals.

IV. 5G NR MMWAVE FOR UAV CAPACITY BOOST

The mmWave spectrum offers wide bandwidth and directional links for high-capacity UAV communications, evaluated through ray-traced urban deployments and system-level simulations.

  • IV. 5G NR MMWAVE FOR UAV CAPACITY BOOST: mmWave provides large bandwidth and highly directional links that can reduce interference and support ambitious 5G aerial use cases.Its directionality helps facilitate coexistence between terrestrial and aerial devices.
  • IV. 5G NR MMWAVE FOR UAV CAPACITY BOOST: Ray tracing with Wireless InSite captures urban aerial coverage because a calibrated mmWave aerial channel model was unavailable.The simulations model angular and power distributions of signal paths together with UAV and base-station antenna patterns.
  • IV. 5G NR MMWAVE FOR UAV CAPACITY BOOST: The urban evaluation uses a 1 km × 1 km 3D section of London and two complementary 28 GHz networks with 400 MHz bandwidth.The networks comprise street-level standard base stations and dedicated rooftop base stations.

1) Coverage with Standard NR mmWave BSs:

Standard NR mmWave cells can provide strong UAV coverage at typical UMi densities through sidelobes and reflections, while dedicated rooftop cells improve robustness in sparse deployments.

  • 1) Coverage with Standard NR mmWave BSs:: At ISDs = 200 m, at least 98% of UAVs at 30 m have SNR > −5 dB, while almost all UAVs at 120 m exceed 15 dB.Coverage improves with altitude despite highly directional, downtilted base-station antennas.
  • 1) Coverage with Standard NR mmWave BSs:: The lower SNR tail is predominantly NLoS, whereas the middle region is LoS-dominated and the upper tail contains strong NLoS paths.Coverage therefore relies substantially on antenna sidelobes combined with reflected paths.
  • 1) Coverage with Standard NR mmWave BSs:: Standard mmWave coverage decays for low-altitude UAVs as ISDs exceed 200 m, requiring progressively higher minimum altitudes for service.At typical UMi densities, sidelobes and strong reflections nevertheless make coverage satisfactory.
  • 2) Coverage Enhancement with Dedicated mmWave BSs:: Higher-altitude UAVs increasingly prefer dedicated rooftop cells as dedicated-cell density rises, while the preference weakens as those cells become sparser.Adding dedicated cells brings LoS-link percentages close to 100% across the considered altitudes.
  • 1) Coverage with Standard NR mmWave BSs:: Larger antenna arrays sustain better SNR and latency, especially at greater UAV–base-station distances.The suburban/rural evaluation compares 2×2 and 4×4 UAV arrays with 4×4 and 8×8 base-station arrays at 28 GHz and 1 GHz bandwidth.

V. UAV-TO-UAV CELLULAR COMMUNICATIONS

UAV-to-UAV cellular communication supports swarms, autonomous flight, aerial relaying, and connectivity extension, but its links and power-control requirements differ from ground D2D.

  • V. UAV-TO-UAV CELLULAR COMMUNICATIONS: U2U links support swarm task distribution, collision avoidance, self-control, and low-latency autonomous operation.Swarm communication is needed because individual UAVs have limited processing capability, payload capacity, and communication range.
  • V. UAV-TO-UAV CELLULAR COMMUNICATIONS: U2U communication enables multi-UAV relaying to extend ground-BS coverage and provide high throughput where aerial cellular service is unreliable.A hovering UAV head can remain well connected to the ground while serving other aerial users.
  • V. UAV-TO-UAV CELLULAR COMMUNICATIONS: Sky D2D reuses four link types—UAV-to-UAV, UAV-to-BS, GUE-to-BS, and GUE-to-UAV—with distinct propagation conditions.UAV-to-UAV links are mostly LoS and depend on UAV altitudes and building heights.
  • V. UAV-TO-UAV CELLULAR COMMUNICATIONS: Fractional power control sets per-PRB transmit power from maximum power, PRB usage, baseline power, and a fraction of large-scale loss compensation.The scheme aims to extend battery life while maintaining received-signal quality and constrained interference.

2) Spectrum Sharing Strategy:

The paper compares underlay and overlay spectrum sharing for coexisting GUE uplink and direct UAV-to-UAV links. Underlay enables broader UAV spectrum access but introduces aerial interference whose effects depend on altitude, power control, and link distance.

  • Spectrum-sharing strategies: Underlay U2U links reuse the GUE spectrum and create co-channel interference, whereas overlay partitions spectrum to prevent cross-tier interference.Underlay aggressiveness is captured by ηu, while overlay gives each communication type non-overlapping spectrum.
  • Altitude effects: Higher UAV altitudes degrade both GUE uplink and U2U performance because more interfering cross-tier links become line-of-sight.The effect is examined for the most aggressive underlay case, where UAVs access the entire GUE uplink spectrum.
  • Power control and distance: Increasing the UAV power-control factor improves U2U performance but reduces GUE uplink performance through stronger aerial interference.Longer U2U distances also reduce both GUE and U2U performance, while high UAV powers increase useful U2U signal power.
  • Overall comparison: Overlay provides the best performance trade-off between direct aerial communication and legacy GUE communication.The comparison considers 5%-worst rates with UAV access restricted to 1 MHz in overlay and extended to 10 MHz in underlay.

D. mmWave U2U Communications

The paper identifies mmWave as a promising route for high-capacity UAV-to-UAV communication, while emphasizing that mobility and wobbling can undermine directional links. It places these links within broader 6G architectures involving additional bandwidth, infrastructure, and network intelligence.

  • D. mmWave U2U Communications: MmWave bandwidth could boost U2U performance, but directional arrays are needed to compensate for higher propagation loss.Directional transmission also reduces interference and supports coexistence between terrestrial and aerial devices.
  • D. mmWave U2U Communications: Double UAV wobbling can significantly affect the optimal mmWave U2U antenna design at 60 GHz.For a 500 m link, larger arrays may increase beam misalignment and worsen performance depending on wobbling and communication distance.
  • 6G motivation: Future air taxis, cargo UAVs, and wirelessly backhauled flying base stations will require sufficient capacity and reliability from one or more 6G paradigm shifts.The paper frames this need alongside autonomous eVTOL connectivity and emerging beyond-5G technologies.
  • Non-terrestrial networking: Non-terrestrial networking integrates UAVs, high-altitude platforms, and satellites as flying connectivity components.The paper presents these platforms as part of prospective ground-air-space architectures.

A. UAV-to-Satellite Communications

The paper presents UAV-satellite communication as a route to resilient coverage, backhaul, and remote IoT connectivity. It also identifies hardware, propagation, mobility, and spectrum-management constraints that limit practical deployment.

  • A. UAV-to-Satellite Communications: NTN combines UAVs, HAPs, and satellites, offering broad line-of-sight coverage and resilience where terrestrial infrastructure is unavailable.Terrestrial-NTN combinations can support BVLoS control across cells spanning tens to hundreds of kilometers.
  • Backhauling to UAV Radio Access Nodes: Satellite backhaul for UAV radio access nodes is challenging because compact terminals must support roughly comparable data demand, with latency favoring LEO satellites.GEO remains possible for latency-tolerant traffic when integrated with terrestrial networks and traffic classification.
  • Aggregators for IoT Services: NTN can facilitate remote-area IoT data collection when the task is not latency-sensitive, although considerable uplink capacity remains necessary.Direct satellite access is generally unsuitable for low-cost, low-form sensors except through LEO.
  • Hardware: Satellite links require specialized antennas that can increase UAV power and weight, while orbit choice trades path loss against field of view, tracking, and Doppler compensation.LEO reduces path loss but moves rapidly; GEO requires more power because of its larger propagation distance.
  • PHY and spectrum: Several satellite technologies remain constrained: mobility-induced channel-state variation affects precoding, inter-orbit carrier aggregation is elusive, and beam hopping may interrupt UAV access.Early active antennas also have limited beam granularity, update rate, and bandwidth, while NOMA gains require practical verification.

3) Protocols:

The paper examines satellite and cell-free architectures as protocol-level directions for extending UAV connectivity. Cell-free MMSE operation substantially improves UAV SINR, while satellite links could fill coverage gaps but face hardware and protocol constraints.

  • 3) Protocols:: Satellite integration can use DVB-compatible architectures or adapt NR waveforms to ground-to-space channel conditions.Waveform adaptation is more challenging because large satellite cells create HARQ-timer, timing-advance, and Doppler-compensation incompatibilities.
  • 3) Protocols:: Satellites could fill ground coverage gaps, support highly mobile UAVs, or backhaul aerial radio access nodes.The paper identifies hardware, PHY, and protocol design as remaining UAV-to-satellite challenges.
  • Cell-free architectures: Cell-free operation coordinates distributed base stations so interference can be classified and routed as useful signal rather than discarded.This architecture distributes spatial degrees of freedom across the network instead of concentrating them in a macrocellular massive-MIMO site.
  • Cell-free architectures: In a UMi UAV-only network, MMSE cell-free processing delivers median SINR gains of 20–30 dB, depending on channel-estimation accuracy.Matched-filter cell-free processing improves worst-case SINR but does not otherwise improve performance; channel-estimation accuracy becomes important under MMSE processing.
  • Cell-free architectures: With 14 GUEs per UAV, the MMSE cell-free advantage remains largely intact despite somewhat fewer cellular outages.The reduced outage is attributed to the lower LoS probability of intercell interferers affecting GUEs.
  • Cell-free architectures: Cell-free architectures still require solutions for scalability, downlink precoding, power allocation, shared computation, altitude dependence, and UAV battery schedules.These issues become especially relevant when cell-free systems move toward C-RAN structures.

IX. ARTIFICIAL INTELLIGENCE TO MODEL AND ENHANCE UAV COMMUNICATIONS

The paper presents AI for two UAV communication needs: accessible aerial channel modeling and mobility management. Neural generative models reproduce channel characteristics, while reinforcement learning reduces handovers through a tunable handover-versus-RSRP trade-off.

  • A. AI for aerial channel modeling: Existing accessible mmWave aerial channel models lack comprehensive statistical descriptions of multipath propagation beyond LoS links.Current 3GPP aerial models are calibrated only for sub-6 GHz frequencies.
  • A. AI for aerial channel modeling: A two-stage neural model predicts LoS, NLoS, or outage and then generates path parameters conditionally on the predicted link state.The resulting model is trained using a large ray-tracing-based urban dataset and evaluated for London.
  • B. AI for UAV Mobility Management: Reinforcement learning is suited to UAV mobility management because network-controlled trajectories make handover decisions intricate and dynamic.The approach targets unnecessary handovers while accounting for observed signal strength.
  • B. AI for UAV Mobility Management: The 5%-worst UAVs lose around 4.5 dB at w_HO/w_RSRP = 5/5 versus the greedy baseline, while w_HO/w_RSRP = 1/9 incurs a 1.8 dB deficit.Across all tested combinations, minimum RSRP remains above −80 dBm.
  • B. AI for UAV Mobility Management: AI-based methods can fill aerial channel-modeling gaps and generalize mobility-management schemes for UAV-aware network design.The paper presents these as distinct demonstrated uses of AI in UAV cellular communications.

X. ASSISTING UAV CELLULAR COMMUNICATIONS WITH RECONFIGURABLE INTELLIGENT SURFACES

Reconfigurable intelligent surfaces could improve UAV link efficiency and reliability by controlling reflections and refractions, while THz links offer exceptional short-range capacity. Both directions remain constrained by modeling, deployment, propagation, and hardware challenges.

  • X. RIS-assisted UAV communications: UAVs may encounter blocked air-to-ground links in complex environments, motivating controllable propagation through RISs.The paper frames RISs as one route, alongside infrastructure densification and space-based connectivity, for improving reliability.
  • X. RIS-assisted UAV communications: RISs use configurable passive scattering elements to control electromagnetic waves through transformations such as reflection, refraction, polarization conversion, and focusing.These transformations can be implemented with low-cost electronic circuits such as PIN diodes or varactors.
  • X. RIS-assisted UAV communications: RIS deployments could provide smart reflections from building façades, smart refractions from billboards or windows, or UAV-carried on-demand coverage.UAV-carried surfaces can support ground-to-air-to-ground reflections with a 360° view.
  • X. RIS-assisted UAV communications: RIS research still needs realistic large-scale deployment optimization and joint optimization of UAV trajectories and RIS signal transformations.The paper also identifies electromagnetic-consistent modeling and practical performance evaluation as open issues.
  • XI. UAV communication at THz frequencies: THz communication can support enormous bandwidths and MIMO transmissions under strict LoS conditions, potentially serving short-range UAV fronthaul.The paper broadly defines THz frequencies as 100 GHz–10 THz.
  • XI. UAV communication at THz frequencies: THz UAV links face missing channel models, molecular absorption, angular sparsity, high path loss, nonflat wavefronts, and bandwidth-related hardware costs.Altitude, trajectories, wind, array wobbling, body blockage, and converter power consumption further constrain operation.
  • XI. UAV communication at THz frequencies: At short and intermediate distances, theoretical THz bit rates are described as truly stupendous even if only a fraction is realized.Figure 28 compares 28 GHz and 140 GHz LoS transmissions using different antenna-array sizes and bandwidths.

XII. CONCLUSION

The conclusion synthesizes the paper’s 5G-to-6G assessment of UAV use cases, requirements, and enabling technologies. It reports concrete benefits from NR enhancements and identifies 6G architectures and technologies that could improve coverage, reliability, mobility, and capacity while leaving substantial open problems.

  • XII. CONCLUSION: The article blends academic and industrial views to examine UAV use cases, requirements, and enabling technologies from 5G through 6G.It presents the topic as one with many remaining fundamental challenges.
  • XII. CONCLUSION: NR enhancements support stringent UAV control and payload-data demands, including beamformed cell selection, mMIMO interference management, mmWave coverage, and UAV-to-UAV links.The conclusion also reports a power-control trade-off between UAV-to-UAV and ground-link performance.
  • XII. CONCLUSION: 6G candidates include NTNs, cell-free architectures, RISs, AI, and THz frequencies for improving UAV coverage, reliability, network design, and capacity.The conclusion connects these technologies to ground-coverage gaps, interference management, aerial channel modeling, mobility management, and LoS MIMO.
  • XII. CONCLUSION: Tailoring these technologies to UAVs remains challenging, motivating open problems and further research.The paper presents future work as necessary for progress toward the fly-and-connect era.
Loading 2105.04842v1…