PaperlayerView exact revision on arXiv

Source-linked AI summary

Accessing From The Sky: A Tutorial on UAV Communications for 5G and Beyond

Yong Zeng, Qingqing Wu, Rui Zhang

arXiv:1903.05289v2eess.SPcs.IT

TL;DR

UAV communications must address LoS-dominant channels, aerial-terrestrial interference, differing CNPC and payload requirements, SWAP limits, and 3D mobility. This tutorial develops models and optimization frameworks, then reviews UAV-assisted and cellular-connected communication paradigms. Height-dependent compensation factors, including αUE = 0.8 for lower-altitude users and αUE = 0.7 above 100 m, provide significant performance gain over a common factor.

  • Problem

    UAV communications face LoS-dominant air-ground channels, severe interference, distinct CNPC and payload requirements, SWAP constraints, and highly mobile 3D trajectories.

  • Method

    The paper presents channel, antenna, energy-consumption, and trajectory-optimization models and reviews UAV-assisted and cellular-connected communication frameworks.

  • Results

    Significant performance gain is attained with height-dependent compensation factors, using αUE = 0.8 for terrestrial and aerial UEs below 100 m and αUE = 0.7 for aerial UEs above 100 m.

  • Takeaways & Limitations

    Existing LTE networks can support initial low-density, low-altitude UAV deployment, but larger-scale deployment requires improved 3D coverage and air-ground interference mitigation.

Abstract

from arXiv · show

Unmanned aerial vehicles (UAVs) have found numerous applications and are expected to bring fertile business opportunities in the next decade. Among various enabling technologies for UAVs, wireless communication is essential and has drawn significantly growing attention in recent years. Compared to the conventional terrestrial communications, UAVs' communications face new challenges due to their high altitude above the ground and great flexibility of movement in the three-dimensional (3D) space. Several critical issues arise, including the line-of-sight (LoS) dominant UAV-ground channels and resultant strong aerial-terrestrial network interference, the distinct communication quality of service (QoS) requirements for UAV control messages versus payload data, the stringent constraints imposed by the size, weight and power (SWAP) limitations of UAVs, as well as the exploitation of the new design degree of freedom (DoF) brought by the highly controllable 3D UAV mobility. In this paper, we give a tutorial overview of the recent advances in UAV communications to address the above issues, with an emphasis on how to integrate UAVs into the forthcoming fifth-generation (5G) and future cellular networks. In particular, we partition our discussions into two promising research and application frameworks of UAV communications, namely UAV-assisted wireless communications and cellular-connected UAVs,where UAVs serve as aerial communication platforms and users, respectively. Furthermore, we point out promising directions for future research and investigation.

I. INTRODUCTION

UAV communication must support safety-critical control links and application-dependent payload data across highly mobile aerial networks. The paper frames direct, satellite, ad-hoc, and cellular networking as candidate technologies, while emphasizing cellular integration for scalable deployment.

  • Communication requirements: Wireless communication supports both safety-critical control and non-payload communication (CNPC) and mission-related payload data for UAV operations.CNPC includes command, air-traffic-control relay, and sense-and-avoid communication, whereas payload links carry data such as aerial images and video.
  • Communication requirements: CNPC links prioritize reliability and low latency, while payload links require application-dependent data rates ranging from several Mbps for FHD video to over 30 Mbps for 4K video.CNPC requirements are similar across UAV types because they serve common safety needs, whereas payload requirements vary by mission.
  • Wireless technologies: The paper compares direct links, satellites, ad-hoc networks, and cellular networks as candidate technologies for reliable air-to-air and air-to-ground communication in 3D space.It presents cellular networking as a promising basis for scalable UAV-ground connectivity because existing and future cellular systems provide broad coverage and advanced backhaul.
  • Wireless technologies: Direct-link communication is simple and inexpensive but is constrained by line-of-sight, limited range, urban blockage, insecure operation, and vulnerability to interference and jamming.The paper states that these limitations prevent direct links from serving as a scalable solution for large-scale UAV deployment.
  • Wireless technologies: Satellite communication offers global coverage and remote-area connectivity but introduces substantial propagation loss and delay, SWAP burdens, and high operating costs.These limitations particularly challenge ultra-reliable, delay-sensitive CNPC and densely deployed consumer applications.

C. The New Paradigm: Integrating UAVs into Cellular Network

The paper frames cellular integration around two paradigms: UAVs as aerial users and UAVs as aerial communication platforms. It surveys their opportunities, challenges, models, and design techniques for future cellular networks.

  • The New Paradigm: Cellular integration treats UAVs either as cellular-connected aerial users or as aerial communication platforms supporting terrestrial networks.These two paradigms organize the paper’s discussion of UAV integration into cellular networks.
  • UAV-Assisted Wireless Communications: UAV-mounted base stations or relays can assist terrestrial communications because compact communication equipment makes aerial deployment feasible.Commercial LTE base stations weighing less than 4 Kg are cited as suitable for UAV mounting with moderate payload.
  • UAV-Assisted Wireless Communications: UAV-assisted communications can support cellular offloading, information dissemination, sensor and IoT data collection, and geographically separated data transfer.These applications exploit UAV mobility and aerial placement to address dynamic wireless traffic demands.
  • Challenges: High altitude, LoS-dominant channels, 3D mobility, and SWAP constraints create distinctive coverage, interference, handover, and endurance challenges.Aerial links can improve coverage and simplify scheduling, but LoS conditions intensify interference and high mobility increases handovers and time-varying backhaul links.
  • Prior Work and Contribution: The tutorial contributes a unified framework for joint UAV trajectory and communication design and surveys techniques for mitigating air-ground interference.It also synthesizes academic and industrial research across cellular-connected and UAV-assisted communication frameworks.
  • Paper Scope: The paper covers channel and energy models, trajectory-communication co-design, UAV-assisted systems, cellular-connected UAVs, field trials, standardization, and interference mitigation.Its organization extends from fundamentals through both major application paradigms and future research directions.

A. Channel Model

The channel-model discussion focuses on GBS-UAV and UAV-GT links, whose elevated propagation conditions require models beyond conventional terrestrial assumptions. It presents free-space, altitude-dependent, and elevation-angle-dependent approaches for representing UAV-ground channels.

  • Channel Model Scope: UAV communications involve GBS-UAV, UAV-GT, and UAV-UAV links, with the tutorial focusing on GBS-UAV and UAV-GT channel models.UAV-UAV links are usually modeled using free-space path loss when communicating over moderate distances in clear airspace.
  • General Channel Model: UAV-ground channels commonly require Rician or Nakagami-m fading models because a line-of-sight component is usually present.This differs from the Rayleigh fading model commonly used for terrestrial small-scale fading.
  • Free-Space Channel Model: The free-space model removes shadowing and small-scale fading, making channel power depend entirely on transmitter-receiver distance and locations.This model is widely used in early offline UAV trajectory-optimization studies because locations make the channel predictable.
  • Free-Space Channel Model: Free-space propagation is reasonable in rural or sufficiently high-altitude settings but is oversimplified for low-altitude urban UAVs.Urban models therefore use altitude- or elevation-angle-dependent parameters, or probabilistic LoS and NLoS modeling.
  • Altitude-Dependent Channel Parameters: Altitude-dependent models generally reduce the path loss exponent as UAV altitude increases, approaching free-space propagation with α = 2 at sufficiently large altitude.The reduction reflects fewer obstacles and less scattering, while GBS-UAV and UAV-GT links may require different fitted parameters.
  • Elevation-Angle-Dependent Channel Parameters: Elevation-angle-dependent models capture propagation changes caused by both UAV altitude and horizontal distance from the ground node.As elevation angle increases, reduced obstruction and scattering compete with increased link distance and reduced GBS antenna gain.

3) Probabilistic LoS Channel Model:

The probabilistic LoS channel model represents UAV-ground propagation by weighting LoS and NLoS conditions according to elevation angle or deployment-dependent altitude and distance. It captures an altitude tradeoff: expected channel power rises initially as LoS becomes more likely, then falls when added path loss dominates.

  • Elevation-angle-dependent probabilistic LoS model: The elevation-angle-dependent model represents LoS probability with a logistic function that increases with elevation angle and approaches one at sufficiently large angles.The NLoS probability is one minus the LoS probability.
  • Elevation-angle-dependent probabilistic LoS model: Expected channel power averages over building randomness and small-scale fading while incorporating NLoS attenuation through a regularized LoS probability.The regularized probability is P̂LoS(θ) = PLoS(θ) + (1 − PLoS(θ))κ.
  • Modeling assumption: The elevation-angle model simplifies NLoS attenuation by assuming a homogeneous κ, although practical κ is random and log-normally distributed.This assumption is explicitly identified as a simplification.
  • Altitude tradeoff: With fixed 2D distance, expected channel power first increases with UAV altitude, then decreases beyond a threshold when increased path loss outweighs the LoS-probability benefit.This altitude tradeoff is used in UAV-mounted BS and relay placement optimization.
  • 3GPP GBS-UAV channel model: The 3GPP GBS-UAV model covers altitudes from 1.5 m to 300 m and specifies LoS probability, path loss, shadowing, and small-scale fading for three deployment scenarios.For intermediate altitudes H1 ≤ H_U ≤ H2, path loss exponent and shadowing standard deviation decrease as altitude increases.
  • 3GPP GBS-UAV channel model: In the 3GPP model, LoS probability depends on 2D GBS-UAV distance and altitude, with terrestrial, intermediate, and 100% LoS regimes separated by thresholds H1 and H2.Suggested H2 values differ by deployment scenario, including 40 m for RMa and 100 m for UMa.

4) Comparison of Different Models:

UAV-ground channel and antenna models offer different tradeoffs between analytical tractability and modeling accuracy across rural, urban, wideband, and directional-antenna settings. The paper also identifies unresolved modeling and channel-estimation needs caused by UAV mobility, orientation, and aerial users’ side-lobe reception.

  • Channel-model comparison: Free-space models are favored for trajectory design because of their simplicity and good approximation in rural environments or at sufficiently high UAV altitude.Urban settings require models that account for more complex propagation conditions.
  • Channel-model comparison: Channel-model selection depends on communication scenario and study purpose, balancing analytical tractability against modeling accuracy.Channel measurements and modeling remain active research areas.
  • Channel-model comparison: Alternative UAV-ground models include 3D geometry-based stochastic MIMO models, two-ray models over water, measurement-based path-loss models, and tapped delay-line wideband models.Measurements have covered L-band and C-band air-ground channels in mountainous, suburban, and near-urban environments.
  • Future directions: Future work includes MIMO and massive-MIMO modeling, mobility- and blade-rotation-induced channel variation, mmWave and wideband models, and efficient channel estimation.Channel-estimation research is especially needed for the specific UAV models used with MIMO or massive MIMO.
  • Antenna-model comparison: UAV-ground communications generally require 3D antenna models that account for both azimuth and elevation angles, unlike conventional predominantly horizontal 2D modeling.UAV orientation and antenna boresight may continuously change during flight.
  • Antenna-model comparison: Fixed-pattern directional antennas can be modeled with deterministic gain functions or synthesized array patterns, while approximations support theoretical analysis.An eight-element ULA example directs its main lobe toward a −10° elevation angle.
  • Antenna-model comparison: The two-lobe antenna approximation may be insufficient for cellular UAV analysis because aerial users can receive desired signals or interference through distinct side lobes.More accurate gain approximations are needed to distinguish the strongest side lobe from other side lobes.

2) UAV MIMO Communications:

UAV communications require antenna and transceiver designs that account for elevation-aware 3D propagation, limited multipath, and stringent SWAP constraints. The section also models propulsion energy across UAV types and speeds to support energy-aware communication design.

  • UAV MIMO Communications: UAV MIMO channels require both azimuth and elevation angles, with channel coefficients represented through transmit and receive array-response vectors.The model uses multipath components with elevation and azimuth angles of arrival and departure.
  • UAV MIMO Communications: 3GPP suggests uniform rectangular arrays at cellular BSs, placing antenna elements along both vertical and horizontal dimensions for MIMO UAV communications.A suggested UMa configuration is (M1, M2, P) = (8, 4, 2).
  • UAV MIMO Communications: UAV MIMO transceivers must reduce RF-chain cost, processing complexity, and energy consumption because massive arrays are difficult to implement under UAV constraints.Analog beamforming, hybrid precoding, and lens arrays are identified as cost-aware architectures; lens MIMO is especially attractive in sparse multipath environments.
  • UAV Energy Consumption Model: Fixed-wing UAVs require infinite propulsion power at zero speed, whereas rotary-wing UAVs can hover with finite power P0 + Pi.This reflects the fixed-wing minimum-forward-speed requirement and rotary-wing hovering capability.
  • UAV Energy Consumption Model: Both UAV types combine parasite and induced power, while rotary-wing UAVs additionally require blade profile power.Parasite power increases cubically with speed, whereas induced power decreases as speed increases.
  • UAV Energy Consumption Model: The maximum-endurance speed is generally nonzero for rotary-wing UAVs because induced power decreases as speed increases.Thus, hovering is not generally the most power-conserving rotary-wing operating condition.
  • UAV Energy Consumption Model: The trajectory energy approximation ignores acceleration-related external work and assumes zero wind, motivating rigorous 3D derivations and flight-experiment validation.The paper identifies wind effects and practical validation as open research needs.

D. UAV Communication Performance Metric

UAV communication evaluation uses conventional link metrics while adding mission completion time and energy consumption for scenarios where flight objectives matter.

  • UAV communications can use SINR, outage or coverage probability, throughput, delay, spectral efficiency, and energy efficiency as performance metrics.

1) SINR:

The SINR framework models how UAV locations affect desired signal power and terrestrial or aerial interference. This distinction is central to outage, throughput, and interference-aware trajectory analysis.

  • 1) SINR: For a transmitting UAV, SINR depends on its desired received signal power, terrestrial interference, aerial interference, and receiver noise.The desired signal and aerial interference vary with UAV locations.
  • 1) SINR: For a receiving UAV, changing its location affects SINR through both desired signal power and terrestrial and aerial interference powers.This is more complicated than the transmitter case, where location affects own-link SINR through desired signal power only.
  • 1) SINR: The desired signal power depends on transmit power, antenna gains, large-scale channel power, and small-scale fading.UAV location enters through the transmit gain, receive gain, and large-scale channel power.
  • 1) SINR: Outage probability for a target SINR threshold incorporates randomness from both time-varying small-scale fading and spatial LoS/NLoS conditions.For fixed UAV locations, both temporal and spatial randomness remain relevant.
  • 1) SINR: Under Gaussian signaling and Gaussian interference and noise, the achievable rate is Rk(Q) = log2(1 + γk(Q)) bps/Hz for each channel realization.Average throughput is then obtained over random channel realizations and can be extended to UAV trajectories.

4) Energy Efficiency:

UAV communication energy efficiency accounts for both information throughput and the propulsion energy required for flight. The paper formulates trajectory and communication variables jointly under physical, mission, and regulatory constraints.

  • Energy-efficiency metrics: Energy efficiency measures reliably communicated information bits per unit energy, including both propulsion and communication energy consumption.Per-link efficiency uses a UAV’s average throughput divided by its propulsion and communication energy; network efficiency aggregates all links.
  • Joint design: The joint optimization selects UAV trajectories and communication variables such as transmit power, bandwidth, scheduling, and beamforming.The utility may represent different performance metrics, while constraints can apply separately to trajectories, communications, or both.
  • Trajectory constraints: UAV mobility is constrained by initial and final locations, speed, acceleration, obstacle avoidance, collision avoidance, no-fly zones, and altitude limits.These constraints reflect aircraft mechanics, mission requirements, and flying regulations.
  • Optimization challenges: The trajectory-and-communication co-design problem is generally difficult because it is non-convex and involves continuous-time variables.Even fixing either trajectory or communication variables may leave a non-convex problem, while continuous time creates infinitely many variables.

III. UAV-ASSISTED WIRELESS COMMUNICATIONS

UAV-assisted wireless communications use UAVs as aerial platforms serving terrestrial users. The framework covers coverage, relaying, information dissemination, placement, mobility, energy efficiency, and learning-based design.

  • Framework and use cases: UAVs can act as aerial base stations, relays, or access points for terrestrial wireless communications.These roles support ubiquitous coverage, connectivity between distant ground users, and information dissemination or collection.
  • Communication models: Representative communication models include relaying, downlink, uplink, multicasting, and multi-UAV interference channels.Systems may also coexist with terrestrial base stations, access points, or relays.
  • Mobility regimes: Research distinguishes quasi-stationary platforms, which emphasize placement, from flying platforms, which exploit trajectory optimization.The practical choice between static and flying UAVs depends on application requirements.
  • Research directions: The section reviews performance analysis, UAV placement, trajectory and communication co-design, energy-efficient communication, and machine-learning-based design.These topics organize the main research directions for UAV-assisted communications.

B. Performance Analysis

Performance analysis evaluates UAV-assisted systems through experiments, simulations, and theory, using models that vary in node placement, system setup, channels, antennas, and mobility.

  • Analysis approaches: UAV-assisted systems are evaluated through field tests, computer simulations, or theoretical analysis using metrics such as coverage, outage probability, and expected throughput.Theoretical analysis predicts expected performance before deployment.
  • Spatial and mobility models: Deterministic models specify UAV locations or trajectories, while stochastic models use point processes for network-level analysis of multiple UAVs.Stochastic geometry must account for 3D deployment, interference, and more sophisticated UAV channel models.
  • Static platforms: As altitude increases, coverage probability may decrease or may first increase and then decrease, depending on the channel and antenna models.One model attributes degradation to diminishing distance differences between desired and interfering links, whereas directional-antenna models produce a non-monotonic trend.
  • Flying platforms: Stochastically moving UAV base stations achieve comparable coverage to static ones with significantly reduced channel average fade duration.The result was obtained for spiral and oval stochastic trajectory processes modeled as a binomial point process at each snapshot.

1) No ULI:

Without user-location information, UAV placement and trajectory design target coverage and communication quality using limited assumptions. Mobility adds a design degree of freedom, but initial path planning and joint optimization remain challenging.

  • UAV placement: With no user-location information, UAV placement is typically optimized to maximize geographic coverage.Representative placement objectives also include covered users, throughput, or minimizing the number of UAVs when additional information is available.
  • Mobility exploitation: Flying UAVs provide an additional design degree of freedom through trajectory optimization compared with terrestrial or quasi-stationary platforms.Three-dimensional airspace offers greater path flexibility than ground mobility constrained by obstacles.
  • Initial path planning: Trajectory optimization jointly considers path planning and speed optimization to determine the route and time spent at locations.TSP and related methods can initialize routes before refined trajectory-and-communication optimization using BCD and SCA.
  • Mobility and channel quality: Moving closer to a ground terminal can improve channel path loss by about 23 dB for both LoS and NLoS cases and average channel power by about 40 dB.The power gain includes the additional benefit of increased LoS probability in the illustrated point-to-point setting.
  • Initial path planning: TSP determines a flying path and serving order when the operation duration is sufficient to reach all ground terminals, while TSPN allows neighborhood visits instead of exact terminal visits.TSPN is useful when reaching the exact location is unnecessary for the communication task.

2) Joint Trajectory and Communication Optimization:

Joint UAV trajectory and communication optimization is difficult because continuous-time trajectories create infinitely many variables and the resulting problem is generally non-convex. The tutorial addresses this through trajectory discretization, alternating optimization, and successive convex approximation.

  • Solution methods: SCA requires a feasible initial trajectory, for which TSP/PDP path planning can provide a starting point, although more general initialization methods remain necessary under varied mobility constraints.
  • The generic problem jointly optimizes UAV trajectory and communication resource allocation, but remains non-convex even when either variable block is fixed.
  • Trajectory discretization: Time discretization divides the operation horizon into equal slots, producing a finite trajectory sequence with linear state-space constraints for location, velocity, and acceleration.
  • Trajectory discretization: Path discretization represents the route with generally unequal line segments and their durations, usually requiring fewer segments than time discretization.
  • Trajectory discretization: A 1000 s hover requires 1000 location variables under 1 s time discretization but only three variables under path discretization.
  • Solution methods: Block coordinate descent alternates communication allocation and trajectory updates, while SCA converts the trajectory subproblem into convex problems with monotonic convergence to at least a KKT solution under mild conditions.

E. Energy-Efficient UAV Communication

Energy-efficient UAV communication must account for propulsion energy alongside communication energy because onboard energy is highly limited. The tutorial examines trajectory-dependent energy–throughput tradeoffs for fixed-wing and rotary-wing UAVs.

  • Existing energy-efficient studies optimize propulsion or communication objectives under trajectory and rate constraints using models tailored to UAV type.
  • UAV propulsion energy is usually much greater than communication energy, fundamentally limiting endurance and communication performance.
  • For a fixed-wing UAV on a circular path, the elevation angle varies with radius as θ(r) = tan^-1(HU/r), linking path geometry to the channel model.
  • As the circular-path radius increases, both radius-dependent numerator and denominator terms decrease, implying an optimal radius in the energy-efficiency expression.
  • Fixed-wing and rotary-wing UAVs require different energy models because their mechanical designs and energy consumption differ fundamentally.
  • Energy-efficient designs trade reduced ground-node transmission energy against increased UAV propulsion energy when the UAV flies closer to ground terminals.

F. UAV-Assisted Communication via Intelligent Learning

Statistical channel models and imperfect environmental knowledge limit guaranteed real-time trajectory optimization. The tutorial therefore surveys city maps, radio maps, and reinforcement learning as ways to learn or adapt to propagation conditions.

  • Statistical channel models are mainly suitable for average analysis and offline optimization, not guaranteed real-time performance under model mismatch and imperfect knowledge.
  • Accurate 3D city maps allow line-of-sight or non-line-of-sight conditions to be inferred from UAV and user locations through methods such as ray tracing.
  • When city maps are unavailable, UAV measurements from known locations can support construction of a radio map for UAV-enabled relaying.
  • Reinforcement learning enables UAVs to learn from direct environmental interaction and has been applied to navigation, anti-jamming, and communication-rate maximization.
  • Q-learning can learn a multiuser UAV trajectory without explicit user-location or channel information by interacting with a 15 by 15 grid environment.

IV. CELLULAR-CONNECTED UAV

Cellular-connected UAVs can use existing LTE infrastructure for initial low-density, low-altitude deployment, but aerial users create stronger interference and coverage challenges as altitude and scale increase. 3GPP therefore identified requirements and mitigation techniques for larger deployments.

  • Higher-altitude aerial users experience stronger downlink interference: detectable base stations increase, while best-cell SINR can be much lower than for ground users at 150 m or 300 m.
  • 3GPP specified maximum aerial-vehicle height of 300 m and maximum horizontal speed of 160 km/h, alongside a GBS-UAV channel model for heights up to 300 m.
  • Interference management: Interference detection can use UE measurements and mobility information or network exchanges involving scheduling data, RSRP, RSRQ, and CSI.
  • Interference management: Suggested uplink mitigation methods include height-dependent power control and full-dimensional MIMO beamforming for flexible azimuth and elevation control.
  • Field measurements and 3GPP investigations indicate that existing LTE can support initial UAV deployment at low density and low altitude without major network changes.
  • 3GPP identified larger-scale deployment needs for ubiquitous 3D coverage, air-ground interference mitigation, and enhanced CNPC and payload communication support.

C. Performance Evaluation

Performance evaluations show that cellular-connected UAVs create severe interference challenges, while 3D beamforming and multi-cell cooperation offer important mitigation mechanisms. Simulations and theoretical analyses examine association, spectral efficiency, coverage, and reliability under aerial-terrestrial coexistence.

  • Performance Evaluation: At high altitude, fixed downtilted antenna patterns can push UAVs toward distant serving cells, whereas 3D beamforming keeps association near the UAV.Flexible elevation-aware beam adjustment focuses signals toward high-altitude UAVs, avoiding nearby-cell antenna nulls and weak sidelobes.
  • Performance Evaluation: Increasing the number of UAVs degrades overall downlink spectral efficiency because aerial users experience stronger interference than ground users.The empirical CDF in Fig. 18 compares achievable sum rates for a network with 15 total aerial and ground users.
  • Performance Evaluation: 3D beamforming significantly improves system spectral efficiency and offers strong potential for interference mitigation with coexisting aerial and ground users.The same observations extend to uplink communication, where UAVs can generate strong interference toward co-channel base stations.
  • Performance Evaluation: Field tests, numerical simulations, and theoretical analyses consistently identify severe air-ground interference in cellular networks supporting aerial users.UAV uplinks interfere with many co-channel base stations, while downlink UAVs receive interference from multiple non-associated base stations.
  • Performance Evaluation: Massive MIMO can dramatically enhance downlink UAV command-and-control reliability through improved interference mitigation.Practical deployment still requires efficient channel and beam training to handle UAV mobility, Doppler effects, and channel phase variation.
  • Performance Evaluation: CoMP and cooperative NOMA exploit cross-cell coordination or decoding to cancel interference and improve cellular-connected UAV performance.Cooperative NOMA extends interference cancellation by allowing sufficiently strong UAV signals to be decoded at selected occupied base stations.

E. QoS-Aware UAV Trajectory Optimization

QoS-aware trajectory optimization treats UAV mobility as a controllable design variable for maintaining cellular connectivity while satisfying mission objectives. The approach is constrained by coverage-map accuracy and by trade-offs between connectivity, interference, energy, and other application requirements.

  • E. QoS-Aware UAV Trajectory Optimization: Controllable UAV mobility provides an additional design degree of freedom for planning trajectories that maintain communication quality.Paths can be designed to avoid cellular coverage holes and preserve connectivity during flight.
  • E. QoS-Aware UAV Trajectory Optimization: A package-delivery trajectory can minimize flight time while maintaining connection to at least one base station under an outage-probability constraint.The coverage region depends on the target SNR threshold γ and tolerable outage value ϵ, so these parameters affect the optimal path.
  • E. QoS-Aware UAV Trajectory Optimization: Trajectory optimization can enforce zero outage or permit bounded cellular disconnection when the disconnected duration stays below a specified threshold.Free-space LoS channels and isotropic antennas reduce coverage regions to circles in one formulation, enabling graph-theoretic and convex-optimization solutions.
  • E. QoS-Aware UAV Trajectory Optimization: Cruise-height control can reduce uplink interference caused by aerial users, but it usually compromises the UAV’s link quality.This exposes a direct trade-off between interference management and communication performance.
  • E. QoS-Aware UAV Trajectory Optimization: Accurate three-dimensional base-station coverage maps are a practical challenge for optimal QoS-aware path planning.Reinforcement learning has been studied to jointly optimize trajectory, cell association, and power control, but machine-learning research in this area remains at an early stage.
  • E. QoS-Aware UAV Trajectory Optimization: UAV trajectory design interacts with broader communication constraints involving security, caching, mmWave links, edge computing, and energy transfer.For computation offloading, hovering above associated base stations is efficient when task-input size is sufficiently large, but propulsion energy can change that conclusion.
Loading 1903.05289v2…