PaperlayerView exact revision on arXiv

Source-linked AI summary

A Vision and Framework for the High Altitude Platform Station (HAPS) Networks of the Future

Gunes Kurt, Mohammad G. Khoshkholgh, Safwan Alfattani, Ahmed Ibrahim, Tasneem S. J. Darwish, Md Sahabul Alam, Halim Yanikomeroglu, Abbas Yongacoglu

arXiv:2007.15088v4cs.ITeess.SP

TL;DR

Growing metropolitan demand exceeds some terrestrial and UAV-based solutions, while HAPS networks face unresolved waveform, heterogeneous-service, and three-dimensional management challenges. This article develops a literature-supported vision and framework, reviewing enabling technologies and management techniques to substantiate future HAPS deployment opportunities and challenges.

  • Problem

    Metropolitan communication demand is increasing beyond some small-cell and UAV-based solutions, while HAPS lack standardized waveforms and require management of heterogeneous service requirements under energy and payload constraints.

  • Method

    The article provides a comprehensive state-of-the-art literature review and framework covering HAPS use cases, energy and payload systems, RSS, radio resource management, physical-layer techniques, handoff, and AI.

  • Results

    The review substantiates a vision of HAPS as a prospective component of future wireless deployments, including metropolitan service, coverage-gap filling, aerial computing, and dense HAPS constellations.

  • Takeaways & Limitations

    HAPS networks offer multiple deployment and technology opportunities, including RSS payloads, event-aware resource management, AI-enabled onboard processing, and FTN-based optical backhauling.

Abstract

from arXiv · show

A High Altitude Platform Station (HAPS) is a network node that operates in the stratosphere at an of altitude around 20 km and is instrumental for providing communication services. Precipitated by technological innovations in the areas of autonomous avionics, array antennas, solar panel efficiency levels, and battery energy densities, and fueled by flourishing industry ecosystems, the HAPS has emerged as an indispensable component of next-generations of wireless networks. In this article, we provide a vision and framework for the HAPS networks of the future supported by a comprehensive and state-of-the-art literature review. We highlight the unrealized potential of HAPS systems and elaborate on their unique ability to serve metropolitan areas. The latest advancements and promising technologies in the HAPS energy and payload systems are discussed. The integration of the emerging Reconfigurable Smart Surface (RSS) technology in the communications payload of HAPS systems for providing a cost-effective deployment is proposed. A detailed overview of the radio resource management in HAPS systems is presented along with synergistic physical layer techniques, including Faster-Than-Nyquist (FTN) signaling. Numerous aspects of handoff management in HAPS systems are described. The notable contributions of Artificial Intelligence (AI) in HAPS, including machine learning in the design, topology management, handoff, and resource allocation aspects are emphasized. The extensive overview of the literature we provide is crucial for substantiating our vision that depicts the expected deployment opportunities and challenges in the next 10 years (next-generation networks), as well as in the subsequent 10 years (next-next-generation networks).

I. INTRODUCTION

The paper presents HAPS as an integral aerial layer in 6G VHetNets and develops a framework for HAPS mega-constellations serving communication, computation, and networking needs. It reviews enabling technologies, resource and mobility management, waveform design, AI, and deployment challenges.

  • 6G VHetNets are envisioned as three-layer architectures integrating satellite, aerial, and terrestrial networks, with HAPS as an integral component.
  • HAPS operate around 20 km in the stratosphere, where quasi-stationary positioning supports ubiquitous connectivity and broader applications than rural access or disaster relief.
  • A HAPS mega-constellation is envisioned to provide high-capacity access, computation offloading, and data analytics to millions of users and devices in suburban and dense urban areas.
  • The proposed framework includes satellite communication and orbital-data functions, UAV swarm management, edge intelligence, and support for URLLC and eMBB services.
  • The article reviews HAPS architectures, use cases, energy and payload technologies, RSS integration, radio resource and interference management, waveform and multiple-access techniques, handoff, network management, and AI/ML.

II. PROMISING USE-CASES FOR HAPS SYSTEMS IN NEXT-GENERATION NETWORKS

HAPS systems are envisioned as complementary network platforms that extend beyond rural connectivity to address metropolitan coverage, capacity, IoT, computation, and diverse application demands. Their large footprints and onboard capabilities support flexible alternatives to terrestrial densification, while introducing challenges for massive access and device energy constraints.

  • II. PROMISING USE-CASES FOR HAPS SYSTEMS IN NEXT-GENERATION NETWORKS: The HAPS vision spans connectivity and computation across urban centers, unsuitable terrain, and multiple space, aerial, and terrestrial network layers.The proposed HAPS-SMBS use cases include IoT services, backhaul, event coverage, aerial-network support, transportation systems, satellite interfaces, and aerial data centers.
  • A. HAPS-Mounted Super Macro Base Station (HAPS-SMBS): HAPS systems are proposed as complementary platforms for urban and suburban networks, where increasing demand makes terrestrial small-cell densification insufficient.The envisioned architecture uses one platform for multiple applications rather than treating HAPS only as rural or disaster-relief infrastructure.
  • A. HAPS-Mounted Super Macro Base Station (HAPS-SMBS): HAPS-SMBS platforms provide larger footprints, quasi-stationary operation, more computation, and better line-of-sight links than UAV-based base stations.They can rapidly supplement terrestrial networks for variable or unpredictable traffic demands and exploit massive MIMO for increased capacity.
  • 1) HAPS-SMBS systems to support IoT services:: HAPS-SMBS systems can collect IoT data and provide reliable, seamless, cost-effective uplinks across large device populations.Direct communication from low-rate IoT devices may use low transmission power, but the broad coverage area creates mx-MTC requirements.

2) HAPS-SMBS systems for backhauling small and isolated BSs:

HAPS-SMBS systems are presented as a backhaul option for small and isolated base stations, using favorable line-of-sight propagation and adaptable RF/FSO technologies. Their benefits are accompanied by weather sensitivity, beam-management demands, and the need to compare deployment costs for temporary capacity support.

  • 2) HAPS-SMBS systems for backhauling small and isolated BSs:: HAPS-SMBS backhaul can use mmWave or FSO links, with high-gain antennas and MIMO beamforming supporting cost-effective wireless connectivity.FSO is generally more robust in clear weather, while mmWave links can provide an alternative when atmospheric conditions degrade optical performance.
  • 2) HAPS-SMBS systems for backhauling small and isolated BSs:: HAPS-to-ground small-cell links can remain nearly line-of-sight dominant over 20–200 km, with pathloss exponent around 2 and moderate shadowing or fading.A rule-of-thumb calculation suggests a 200 km HAPS link may provide average power gain comparable to a macro base station 1000 meters away.
  • 2) HAPS-SMBS systems for backhauling small and isolated BSs:: Rain and cloud absorption can impose additional mmWave loss, which HAPS systems may compensate for through higher transmission power and directional antenna gain.The loss is described as proportional to 10cr/H, where H is altitude, r is ground-to-HAPS distance, and c captures rain or cloud effects.
  • 2) HAPS-SMBS systems for backhauling small and isolated BSs:: FSO backhaul quality can substantially deteriorate in cloudy, rainy, or foggy conditions, motivating hybrid RF/FSO solutions and automatic technology switching.Sophisticated resource allocation and 3D beamforming may also be necessary because mmWave can have lower spectral efficiency than FSO.
  • 3) HAPS-SMBS systems for covering unplanned events and filling coverage gaps:: HAPS-SMBS systems can opportunistically support temporary events such as flash crowds, but deployment costs must be compared with terrestrial over-engineering.These scenarios can require high data rates while avoiding serious losses from congestion or inadequate connectivity.

6) HAPS-SMBS systems for supporting and managing aerial networks:

HAPS-SMBS systems are positioned as wide-area computational and connectivity platforms for managing aerial networks, UAV traffic, cargo drones, transportation systems, and satellite interfaces. Their deployment depends on reliable low-latency links, onboard processing, satellite-link design, and coordinated regulation.

  • 6) HAPS-SMBS systems for supporting and managing aerial networks:: HAPS-SMBS systems can enhance UAV networks by providing powerful processors, broad data collection, and lower dependence on overcrowded terrestrial stations.Their larger coverage area supports aerial-network data collection, while the text reports lower interference effects than ground base stations for this offloading use case.
  • 6) HAPS-SMBS systems for supporting and managing aerial networks:: Machine learning can enable HAPS-SMBS systems to control and manage UAV networks intelligently with minimum dependence on terrestrial-based control.This requires seamless UAV connectivity with sufficient reliability, coverage, and relatively low latency.
  • 6) HAPS-SMBS systems for supporting and managing aerial networks:: HAPS-SMBS systems can support cargo-drone operations through highly reliable, low-latency communication and computational platforms for path planning and navigation.The proposed operation may combine radio-based and vision-based solutions across large geographical areas.
  • 6) HAPS-SMBS systems for supporting and managing aerial networks:: A HAPS-SMBS can act as an interface between LEO satellites and aerial or terrestrial networks, using its wide footprint to manage frequent satellite handoffs.The architecture includes user-to-HAPS RF links and HAPS-to-LEO links for which FSO is identified as a better choice.
  • A. Aviation Regulations: Large-scale HAPS deployment remains constrained by nationally varying aviation rules and the need for international licensing and operations frameworks.ITU-R regulates spectrum, ICAO governs safety and civil-aviation relations, and national authorities control airspace licensing.

B. Spectrum Regulations

HAPS spectrum regulation specifies substantial bandwidth needs for ground-to-HAPS and HAPS-to-ground links, within broader network architectures connecting airborne, space, and terrestrial components.

  • Spectrum requirements: 396 MHz to 2969 MHz is needed for ground-to-HAPS links, while 324 MHz to 1505 MHz is necessary for HAPS-to-ground links.These bandwidth ranges were determined through ITU technical investigations.
  • Network architecture: HAPS networks can connect HAPS constellations, satellite layers, LAPS nodes, UAV users, and terrestrial networks.The system includes onboard subsystems and non-terrestrial communication nodes, alongside control stations, gateways, terrestrial base stations, and users.
  • Ground infrastructure: Control stations manage HAPS communication operations, coordinate links and resources, and remotely monitor position and direction.They can also handle takeoff and landing while optimizing antenna efficiency and performance.
  • Ground infrastructure: Communications gateways connect HAPS platforms to core networks through wired backhaul and may support direct or gateway-mediated terrestrial communication.Control stations and gateways may be colocated or separately located.

B. Types of HAPS and Related Projects

HAPS platforms are classified by crew status and lifting principle, with platform choice determined by mission needs, deployment conditions, payload, energy, and endurance trade-offs.

  • Manned and unmanned platforms: Unmanned HAPS are preferred for communications because prolonged operations are difficult for human pilots in the stratospheric environment.Manned platforms have mainly served meteorological, scientific, military, and some telecommunications purposes.
  • Lifting principles: Aerostatic HAPS float through buoyancy, whereas aerodynamic HAPS generate lift through movement, with aerostatic platforms including balloons and airships.These categories differ in their underlying physical lifting principles.
  • Mission trade-offs: Aerodynamic HAPS suit emergencies through lower deployment costs, flexible takeoff and landing, and mobility control.Their reduced deployment cost makes them preferable for unplanned events or emergency situations.
  • Mission trade-offs: Aerostatic HAPS suit longer-term applications through larger payload capacities and stronger energy-generation capabilities.Their station-keeping becomes more difficult under strong winds and turbulent conditions.
  • Related projects: Hybrid HAPS combining aerostatic and aerodynamic advantages may be needed for near-future applications.Loon and HAPSMobile formed a long-term strategic relationship to advance both platform types.
  • Energy considerations: Energy management affects flight duration and deployment costs, making it essential for prolonged, feasible, and cost-effective communications operations.Solar energy is widely considered because HAPS operate above clouds and can accommodate large solar panels.

2) HAPS Energy Consumption:

HAPS energy consumption spans flight control and communications payloads, while passive RSS is proposed to reduce payload energy and deployment costs in relay-oriented scenarios.

  • Energy consumption: Flight-control energy covers stability, propulsion, altitude, and direction control, and depends on platform type, weight, and size.Aerodynamic platforms require continuous circular movement, increasing flight-system energy demands.
  • Communications payload: HAPS payloads may operate as relay stations or full base stations, with HAPS-BS platforms requiring more processing capability and energy.Active payloads include antennas, transponders, amplifiers, converters, processors, and filters.
  • Passive HAPS-RSS: Passive RSS can provide a cost-effective HAPS deployment when advanced communication, caching, and computing functions would not be profitable.This is especially relevant when the HAPS mainly relays signals or serves limited numbers of remote users.
  • Passive HAPS-RSS: RSS manipulates incident-wave phases and directions through lightweight metasurfaces, enabling controlled reflection and refraction.Digital metamaterials allow electromagnetic-wave manipulation to be digitally controlled.
  • Passive HAPS-RSS: HAPS-RSS supports rural backhaul to gateways and inter-HAPS links, with configurations sent by a ground control station.The proposed integration coats or equips HAPS with RSS for these communication scenarios.
  • Passive HAPS-RSS: 40% more energy efficiency is reported for RSS-assisted communication than relay-assisted communication.Each reflector-unit configuration consumed 0.33 mW in a reported experiment, while reduced weight and energy use may prolong flight duration.
  • Passive HAPS-RSS: Passive HAPS-RSS can support cost-effective multi-hop inter-HAPS links because its controller requires less energy than active relaying.This is identified as relevant for energy-efficient HAPS constellations.

2) Non-Geometric Stochastic Models:

HAPS channel and system studies use stochastic propagation models alongside performance, resource-management, and waveform analyses to address reliability and efficiency across diverse links.

  • Non-geometric stochastic models: 2 GHz HAPS channel modeling incorporated terrestrial multipath fading while neglecting rain attenuation because it was negligible in that band.The model placed transmitter and receiver at ellipsoid foci with uniformly positioned scatterers.
  • Non-geometric stochastic models: Semi-Markov and tapped-delay-line models represent transitions among line-of-sight, slight-shadowing, and total-obstruction propagation conditions.These state switches are used to study HAPS channel error performance.
  • Non-geometric stochastic models: Markov-based models capture the appearance, disappearance, and spatial correlation of LOS and multipath components in multi-user HAPS environments.Continuous-time Markov processes and birth-death processes support non-stationary channel descriptions.
  • Non-geometric stochastic models: Air-to-air HAPS models include horizontal and vertical mobility and derive time-frequency correlation and Doppler characteristics.The reported results highlight the importance of considering vertical movement.
  • FSO links: FSO communications are investigated for HAPS-to-HAPS connectivity and backhaul because they offer cost-effective, license-free, high-bandwidth links.Multi-hop communication and relaying are suggested to address beam wandering, pointing errors, and atmospheric sensitivity.
  • Radio resource management: Radio resource management targets spectral efficiency, energy efficiency, user admission, service continuity, heterogeneous rates, and fairness.HAPS-SMBS designs consider blocking probability, dropping probability, and resource allocation across service types.
  • Radio resource management: HAPS-SMBS backhaul requires joint power allocation and association, with further research needed for large HAPS-SMBS constellations.The design includes associating small cells or isolated base stations with HAPS-SMBS nodes.
  • Physical-layer design: Physical-layer design determines transmission reliability and spectral efficiency through link-level channel, bit, and waveform choices.Reliability is measured using BER or outage probability.

A. Power Control/Allocation and Interference Management in HAPS systems

HAPS power-control research addresses admission, mobility, multicasting, and interference-related resource decisions for users with differing QoS requirements. The literature also considers coordinated HAPS–terrestrial operation and handoff-sensitive allocation.

  • Power Control and Admission: Power control manages SINR and resource constraints while supporting admission decisions for users with different QoS requirements.Earlier HAPS studies used power allocation for call admission control, maximizing Grade of Service through dropping and blocking probabilities under SINR and power constraints.
  • Access Technologies: The literature includes power-control frameworks developed for WCDMA that may be adapted, with minor modifications, to multicarrier-CDMA or power-domain NOMA.These schemes primarily date from the early 2000s and reflect the air-interface technologies studied at that time.
  • HAPS–Terrestrial Coordination: A hierarchical HAPS–terrestrial system combines overflow and speed-sensitive strategies to direct users in overlapping coverage areas to the appropriate layer.The HAPS provides SMBS coverage while terrestrial cellular towers provide macro-cell coverage.
  • Mobility and Handoff: Mobility-aware admission schemes use received serving- and neighboring-cell signal measurements to reduce handoff call dropping in standalone HAPS systems.One approach tracks serving and next-strongest pilot-channel SIR values and incorporates user speed and direction.
  • Multicasting: Multicast resource allocation selects terrestrial and/or HAPS channels while preserving QoS for unicast traffic.This work targets efficient allocation of transmission resources to multicast traffic streams in integrated cellular and HAPS systems.

3) Joint Radio and Computational Power Management:

Joint radio and computational management extends HAPS resource allocation to aerial edge computing, channel assignment, spectrum sharing, placement, and self-organizing network operation. These decisions must accommodate platform motion, heterogeneous access, interference, and changing user demands.

  • Joint Radio and Computational Power Management: Aerial edge-computing systems jointly manage radio resources with onboard computation power and time consumption.The associated decisions include user association, service sequence, and task partitioning for time-varying computational tasks.
  • Joint Radio and Computational Power Management: Federated learning with an SVM proactively determines user association before service sequencing and task allocation minimize weighted energy and time consumption.
  • Channel Allocation: Dynamic channel allocation exploits multi-user and multi-HAPS diversity by assigning sub-channels according to differing channel attenuation.Attenuation reflects path loss, shadowing, and fast fading experienced by each user.
  • Mobility-Aware Allocation: HAPS horizontal motion caused by stratospheric crosswinds can trigger handoffs for stationary edge users, motivating movement-aware reservation and priority queuing.Other work considers AI-based wireless channel allocation in response to HAPS mobility.
  • Heterogeneous HAPS Access: Two-HAPS systems distinguish users with access to both nodes from users restricted to one node by physical antenna constraints.Smart or steerable antennas provide full access, whereas fixed antennas can limit HAPS choice and potentially reduce QoS.
  • Self-Organizing Networks: HAPS self-organizing networks require self-configuration, optimization, and healing because aerial element positions and user requirements can change over time.Layered aerial architectures formulate placement as a linear binary program to optimize objectives such as served users, achievable rate, or fairness.

1) Adaptive Cell Shaping:

Adaptive HAPS antenna and cell design aims to improve edge performance, limit interference, and manage beam steering and handoffs. Related approaches include irregular cells, coordinated multipoint, massive MIMO, and emerging waveform designs.

  • Adaptive Cell Shaping: Practical antenna beams deviate from ideal uniform in-cell illumination and zero out-of-cell power, especially at millimeter-wave frequencies.Low side-lobes are desirable, while insufficient directivity can spill power outside the cell and increase inter-cell interference.
  • Adaptive Cell Shaping: Elliptical beams have demonstrated superior optimized cell-edge power compared with circular beam patterns when RF link budgets are marginal.The cited formulation optimizes directivity to maximize received power at the cell edge rather than defining cells solely by half-power beamwidth.
  • Adaptive Cell Shaping: A four-actuator beam-steering solution is simpler than individually mounting each aperture antenna on gimbals, but it requires many handoffs.The complexity advantage becomes especially relevant when a platform serves more than 100 cells.
  • Adaptive Cell Shaping: Irregular cell shapes adapt to user distribution and behavior while aiming to limit co-channel interference and reduce handoff and location-update signaling.The cited simulations showed that cells with irregular shapes can be formed.
  • Multipoint Transmission: User-centric JT-CoMP extends coordinated multipoint to HAPS architectures using phased-array beams and pooled virtual base-station equipment.The approach is intended to address cell-edge interference and replace multiple terrestrial cell sites.
  • Massive MIMO: Massive MIMO uses large antenna arrays for diversity, interference-oriented beamforming, and spatial multiplexing, with reported gains for HAPS user grouping and beamforming.The cited numerical study reported that its method outperformed schemes based on the channel correlation matrix.
  • Waveform Design: HAPS lacks a standardized waveform structure, so candidate access and inter-platform waveforms may draw on technologies under investigation for 6G.The paper notes that no active HAPS waveform research had been reported to the authors’ knowledge at the time of discussion.

1) Filter Bank MultiCarrier (FBMC) Scheme:

The section surveys waveform and multiple-access options for HAPS systems, emphasizing spectral efficiency, multiuser support, and handoff challenges shaped by moving platforms and changing footprints.

  • 1) Filter Bank MultiCarrier (FBMC) Scheme:: FBMC shapes individual sub-carriers for flexible spectrum use but incurs long filters, large symbol durations, and high MIMO detection complexity.These drawbacks can limit low-latency and short-burst machine-type communications.
  • 2) Faster-than-Nyquist Signaling:: FTN increases transmission rate beyond Nyquist by intentionally introducing controllable inter-symbol interference, with about 25% higher data rate at the cost of receiver complexity.For binary sinc pulses, Mazo’s result preserves minimum Euclidean distance when τ ≥0.802 with optimal MLSE detection.
  • 2) Faster-than-Nyquist Signaling:: In HAPS-SMBS uplinks, FTN can increase IoT bits/sec/Hertz without additional SNR or transmission errors despite power and bandwidth constraints.The trade-off is additional receiver complexity, which the passage says HAPS receivers can accommodate through substantial onboard computational power.
  • 3) Spectrally Efficient Frequency Division Multiplexing (SEFDM):: SEFDM packs sub-carriers more closely than orthogonal multiplexing, improving spectrum and energy efficiency through controlled inter-carrier interference.Its sub-carrier selection and spacing for HAPS access links remain open research issues dependent on application type.
  • 4) Non-Orthogonal Multiple Access:: NOMA multiplexes users in the power domain and uses successive interference cancellation, offering capacity-region benefits and a capacity–fairness trade-off over orthogonal access.MIMO-NOMA extends this approach by clustering users per beam, supporting more users and massive connectivity.
  • VII. HANDOFF MANAGEMENT IN HAPS NETWORKS: HAPS handoff management must account for changing footprints caused by platform disturbances, because terrestrial fixed-threshold algorithms can cause ping-pong handoffs or delayed switching.Simple circular or hexagonal coverage models also overlook location-dependent path loss in high-altitude HAPS systems.

VIII. HAPS NETWORK MANAGEMENT AND COMPUTATIONAL ROLE

Future HAPS networks require autonomous, coordinated management to handle multi-platform coverage, resource allocation, and aerial edge-computing services. Softwarization and intelligent control are presented as mechanisms for self-organizing operations.

  • Network management: Multiple HAPS deployments increase operational complexity and require autonomous control and coordination among platforms.Two to four ground-based crew members currently oversee mission planning, flight control, sensors, and data assessment.
  • Network management: Swarm Intelligence converges faster and more stably, whereas Reinforcement Learning achieves higher peak coverage but can produce coverage dips.RL also offers coordination resilience but depends on a properly designed reward function for suitable convergence.
  • Network management: HAPS network softwarization uses SDN, NFV, and network slicing to support automated, self-organized control and flexible service deployment.SDN separates data and control planes, while NFV separates network functions from physical devices.
  • Network management: HAPS-based controllers can exploit wide coverage and relative stability to reduce aerial-network configuration-update delays caused by varying links.The approach places software-defined aerial-network control planes on HAPS platforms.
  • Network management: Hierarchical orchestration can coordinate network segments through orchestrators and a hyperstrator for overall resource allocation.This model is identified as relevant to envisioned HAPS-SMBS use cases.
  • Computational role: HAPS can provide a large aerial edge and distributed data-center capability for ultra-reliable, low-latency services.Future management must support reliable collaboration, distributed computation and storage, and intelligent task scheduling.

IX. THE ROLE OF AI IN HAPS SYSTEMS

AI is positioned as both an onboard capability and a management approach for future HAPS networks. The literature covers resource-constrained hardware, learning-based optimization, task planning, and direct execution of AI workloads on HAPS.

  • AI hardware: Resource-constrained MCUs and FPGAs are being developed with embedded GPU, DSP, and machine-learning capabilities for HAPS deployment.Examples include ARM Helium technology and sensors with embedded machine-learning cores.
  • Learning methods: Deep neural networks use offline training and online execution, while reinforcement learning selects actions through iterative interaction with environment states.Deep neural networks can make decisions even for some states not encountered during offline training.
  • AI-enabled networks: AI can orchestrate HAPS systems, enable aerial edge computing, and support collaborative distributed machine learning using data collected across satellite and terrestrial layers.HAPS may act as aerial data centers and divide complex learning tasks across multiple machines.
  • AI-enabled networks: Future HAPS networks will require AI to manage multiple heterogeneous platforms rather than the single or small deployments emphasized by current studies.The expected scale increases the need for managing, controlling, and operating diverse HAPS systems.
  • Applications: Existing studies apply artificial immune algorithms, neural networks, particle-swarm optimization, and Stackelberg games to constellation, handoff, beamforming, and offloading problems.These methods address capacity-per-cost, frequent handoffs, side-lobe suppression, and UAV computation offloading.
  • AI hardware: A commercial off-the-shelf neural-network accelerator was successfully tested on a HAPS, supporting onboard execution of nontrivial AI tasks.Onboard processing can reduce transmitted data and speed analysis of dynamic environments.

X. OPEN ISSUES

Open issues span near-term adaptation of terrestrial technologies and longer-term redesign for interacting HAPS and satellite mega-constellations. Energy, payload, spectrum, and platform trade-offs remain central constraints.

  • Next-generation challenges: Next-generation challenges involve adapting established terrestrial technologies such as massive MIMO and mmWave to HAPS-specific channels and energy constraints.Additional work is needed for limited channel degrees of freedom, restricted transmission energy, and detection without channel statistics.
  • Next-next-generation challenges: Next-next-generation networks may require disruptive designs for interactions between HAPS mega-constellations and satellite mega-constellations.Limited public knowledge of satellite mega-constellation technologies makes HAPS design more challenging.
  • Spectrum: HAPS spectrum planning must protect radio astronomy from unintended interference when unlicensed ISM bands are used.Interference concerns were substantiated during Google’s Project Loon tests in Oceania.
  • Energy and payload: Different HAPS payload capacities and energy profiles create trade-offs among platform type, cost, performance, and flight endurance.These trade-offs become more important when HAPS provide broader data-center functionality beyond conventional communications.
  • Energy and payload: RSS can reduce payload weight and energy consumption, but allocating more surface area to RSS reduces solar-panel area and absorbed solar energy.Higher reflected-signal directionality and spectral efficiency require more RSS surface area.

3) PHY and Related Cross Layer Design:

HAPS physical-layer design remains open because aerial channels, FSO nonlinearities, interference, energy limits, and heterogeneous service requirements differ from terrestrial assumptions. The paper points to waveform, cross-layer, resource-management, and intelligent deployment research directions.

  • PHY and Related Cross Layer Design: No PHY waveform has yet been specified for HAPS, so candidate waveforms and pulse-shaping filters require simulation and system-level analysis for mmWave channels.The analysis must account for HAPS propagation and channel fading characteristics.
  • PHY and Related Cross Layer Design: FSO-link nonlinearities make high-order modulation difficult, motivating single-carrier FTN and suitable detectors for BASK-based signaling.The proposed direction targets higher spectral efficiency for inter-platform, LEO/HAPS, and backhaul links.
  • PHY and Related Cross Layer Design: FTN and SEFDM require joint design of power allocation, channel allocation, and acceleration or squeezing parameters because spectral-efficiency gains can increase ISI.For single-carrier FTN, decreasing the acceleration parameter increases spectral efficiency but degrades performance through larger ISI.
  • Radio Resource Management: HAPS RRM still needs low-overhead methods that reconcile model-based optimization performance with real-time implementation requirements.DNN-based approximations have demonstrated orders-of-magnitude computational speedups over state-of-the-art optimization-based power allocation.
  • Radio Resource Management: RRM formulations should jointly address heterogeneous QoS requirements, renewable energy constraints, and power management across access, inter-platform, and backhaul links.URLLC, broadband, and mMTC services impose substantially different requirements.
  • Channel modeling: HAPS-to-LAPS and HAPS-to-satellite channel models and performance evaluation remain insufficiently developed.The literature lacks a universally agreed, practical, and substantiated channel model for evaluation.
  • Networks Management of HAPS Systems: HAPS deployment and management must account for limited degrees of freedom, correlated channels, interference, energy, payload, and changing platform responsibilities.Self-organizing control is proposed for optimizing resources and deployment as stations evolve from communication platforms toward computation platforms.

9) Handoff Management in HAPS Networks:

Future HAPS networks make handoff management substantially more complex because they may span multiple layers, contain hundreds of stations, support fast-moving users, and use challenging mmWave links. The section therefore points toward intelligent, self-adaptive coordination alongside broader network, computing, and security mechanisms.

  • Handoff complexity: Global HAPS mega-constellations with multiple layers and hundreds of stations exceed the realistic scope of existing simple handoff scenarios.Conventional handoff approaches are described as inefficient for such complicated networks.
  • Handoff complexity: Future handoff management must address both Layer 2 radio-cell transitions and Layer 3 IPv6 mobility procedures in all-IP HAPS networks.The Layer 3 procedures include configuring, registering, and rerouting through a new IPv6 address.
  • Handoff complexity: Rapidly moving cars, trains, and aerial vehicles require handoff solutions that support time-sensitive applications.HAPS coverage is expected to include network entities moving at high speeds, not only smartphone users.
  • Adaptive management: Dynamic, self-adaptive inter- and intra-HAPS handoff is required, with dynamic beamforming proposed to reduce handoff frequency for many users.The passage also identifies transmitted-power minimization or capacity maximization as design objectives needing revision.
  • Adaptive management: Conventional 5G handoff procedures may fail to meet beyond-5G latency requirements because their three-way handshake introduces propagation delay.The passage also notes that mmWave links are vulnerable to atmospheric absorption and weather conditions.
  • Security and integration: HAPS security must cover compromised nodes, communication links, computing platforms, privacy, and autonomous-platform hijacking risks.A compromised HAPS could affect communications across its enormous footprint, while hijacking could endanger nearby aircraft.

2) HAPS Mega-Constellation:

HAPS mega-constellations are presented as a dynamic, heterogeneous infrastructure for extending satellite coverage and supporting communication, computation, caching, and intelligence services. Realizing this vision requires adaptive optimization, resource coordination, and further investigation of technical and system-level challenges.

  • Motivation and architecture: Satellite mega-constellations face an economic requirement to deliver Internet access faster than fiber-optic networks, while laser-link technology remains in its infancy.The passage identifies laser links as a possible route toward that performance goal.
  • Motivation and architecture: HAPS can extend satellite coverage by boosting, combining, and jointly transmitting satellite signals in weak boundary cells and high-interference regions.Their higher computation and communication capabilities support this role.
  • Network dynamics: Intermittent station recharging and heterogeneous platform behaviors make preserving coverage during network reconfiguration a central challenge.Aerostatic stations may remain quasi-stationary, whereas aerodynamic platforms must keep moving.
  • Network dynamics: Meta-learning is proposed to help HAPS networks adapt across changing optimization tasks such as routing, coverage, backhauling, resource allocation, and computation offloading.The approach learns underlying optimization structures from several prominent tasks rather than repeatedly solving each problem from scratch.
  • Network services: Future HAPS systems must support communication for both data exchange and intelligence across diverse devices while meeting demanding 6G connectivity objectives.The shared communication resources serve two disjoint purposes: data communication and communication for intelligence.
  • Framework and outlook: The proposed HAPS framework covers mega-constellations, SMBS deployment, RSS payloads, RRM, physical-layer techniques, mobility management, and AI enablers.The article frames these topics as part of a forward-looking vision spanning next-generation and next-next-generation challenges.
  • Framework and outlook: Open issues include insufficient HAPS channel knowledge, restricted transmission energy, and detection without available channel statistics or models.Massive MIMO and mmWave use in HAPS systems also requires further investigation despite ongoing terrestrial research.
Loading 2007.15088v4…