High Altitude Platforms (HAPs)

Updated 28 January 2026

High Altitude Platforms (HAPs) are stratospheric, quasi-stationary vehicles that provide wide-area, low-latency communications, effectively bridging terrestrial base stations and satellites.
They feature diverse platform types—such as aerostatic, aerodynamic, and hybrid systems—with regenerative, transparent, or RIS payloads to meet varied data and energy needs in modern networks.
By integrating solar power, advanced mMIMO, and AI-driven control, HAPs enhance energy efficiency and resource allocation, supporting sustainable 6G and vertical heterogeneous networks.

High Altitude Platforms (HAPs) — often more precisely High-Altitude Platform Stations (HAPS)—are stratospheric, quasi-stationary vehicles equipped with communication and/or remote sensing payloads. Operating typical altitudes of 17–25 km, they form a unique non-terrestrial node class that bridges the architectural and performance gap between terrestrial base stations and satellites. This layer provides wide-area LoS coverage with low round-trip delay, high aggregate capacity, energy efficiency via solar power, and is increasingly fundamental to the design of energy-aware, flexible, and sustainable 6G networks and vertical heterogeneous networks (vHetNets).

1. System Architectures and Platform Taxonomy

HAPS encompass diverse platform realizations and payload architectures, characterized chiefly by their aerodynamic principles, payload scaling, and station-keeping method. The main types are:

Platform Type	Lift Mechanism	Endurance	Mobility	Payload
Aerostatic (Airship)	Buoyancy	Months	Low (horizons)	High (100s kg)
Aerodynamic (UAV)	Wings	Weeks–months (solar)	High	Moderate
Hybrid	Buoyancy+Wings	Weeks–months	Moderate	Moderate

Typical operational altitudes are 20–25 km, above commercial airspace and weather systems, yielding circular ground footprints of 30–500 km radius (Svistunov et al., 22 Oct 2025, Xing et al., 2021, Alfattani et al., 2022, Chu et al., 2020). HAPS payload architectures are categorized as:

Regenerative (Base-Station) Payloads: Full digital-baseband processing (MAC/PHY), RF chains, massive MIMO, edge-computing/caching servers.
Transparent (Relay/Bent-Pipe) Payloads: RF-only frequency-converting relays, lower SWaP, no onboard computation.
Reconfigurable Intelligent Surface (RIS) Payloads: Large arrays of passive reflecting elements implementing beam steering with minimal active electronics (Alfattani et al., 2022).

The overlay of HAPS on terrestrial RANs can be super-macro (SMBS), relay, or RIS-based, supporting flexible switching according to traffic and energy efficiency metrics (Alfattani et al., 2022).

2. Communications and Link Budget Principles

HAPS-centric networks exploit predominately LoS links, drastically reducing link path-loss exponents (from ≈4 in NLoS urban macrocell to ≈2 in HAPS LoS) (Kement et al., 2022, Xing et al., 2021).

The dominant propagation law is free-space: $PL_{fs}(d,f) = 20 \log_{10}\left(\frac{4\pi f d}{c}\right)\ \text{[dB]}$ with $d$ the slant range (typically $\approx h / \sin\theta$ ), $f$ frequency, and $c$ speed of light. For $f=2$ GHz, $d=20$ km yields $PL_{fs}\approx124$ dB. Atmospheric attenuation (rain/fog) is minimal in the stratospheric segment, but can become relevant in FSO/THz feeder links or under adverse weather (Elkhazraji et al., 8 Nov 2025, Chu et al., 2020).

Capacity is determined by the per-link bandwidth $B$ and SINR: $R_i = B\log_2(1+\text{SINR}_i)$ HAPS with mMIMO arrays can form $N_{\text{beam}}$ directional beams, dynamically sectorizing their coverage (Kement et al., 2022, Javed et al., 2024).

Backhaul is provided from HAPS to the ground via mmWave, sub-6 GHz, or FSO/THz links (Svistunov et al., 22 Oct 2025, Elkhazraji et al., 8 Nov 2025, Xing et al., 2021). HAPS–HAPS mesh (FSO/mmWave) and HAPS–satellite relays also play roles in NTN integration (Lou et al., 2023, Svistunov et al., 22 Oct 2025).

3. Energy Model, Solar Harvesting, and Green Computing

HAPS sustainability arises from a power subsystem integrating high-efficiency photovoltaic arrays (η≈0.20–0.40, $A_{PV}\sim8000$ m²), batteries, and optionally fuel cells (Kement et al., 2022, Abderrahim et al., 2023, Abderrahim et al., 2023). In the steady-state,

$P_{\text{solar}}(t) = \eta_{\text{panel}}\,A_{\text{panel}}\,I_{\text{sun}}(t)$

where $I_{\text{sun}}$ is the stratospheric irradiance. Nighttime operation is sustained by batteries ( $E_b$ ), with the instantaneous energy balance

$E_b(t+\Delta t) = E_b(t) + [P_{\text{solar}}(t) - P_{\text{total}}(t)]\Delta t$

Total load includes communications ( $P_{\text{comms}}$ ), payload processing ( $P_{\text{payload}}$ ), and flight control ( $P_{\text{ctrl}}$ ).

Stratospheric cold ( $-50$ °C – $-15$ °C) yields up to 6–7 $\times$ reduction in data center cooling energy compared to terrestrial implementations. This supports "flying data centers" capable of ~12–14% energy savings with a single HAPS and up to 36% for constellations (Abderrahim et al., 2023, Abderrahim et al., 2023).

Power consumption breakdown and adaptive switching among SMBS/Relay/RIS payload modes extend loitering time by 20–50% in low-power settings (Alfattani et al., 2022). HAPS energy model must be co-optimized across flight dynamics, harvesting, and mission planning (Javed et al., 2022).

4. Network Integration: Terrestrial, Non-Terrestrial, and vHetNet

HAPS function as a critical "stratospheric glue" in multi-tier networks, enabling:

Overlay in 6G RANs: HAPS SMBS overlays urban/dense terrestrial networks, providing rapid "vertical" scaling during demand peaks, with capacity utilization $\uparrow$ and network power draw $\downarrow$ against traditional RAN densification (e.g., 71% vs 31% utilization; 140.6 kW vs. 314.5 kW total power) (Kement et al., 2022, Song et al., 2023).
Non-Terrestrial Network (NTN) Nexus: HAPS interconnect satellites, UAV FANETs, and terrestrial nodes via multi-hop FSO/mmWave, with distinct gains in coverage, energy efficiency, and latency (e.g., inter-HAPS FSO mesh achieves $\ll 1$ ms E2E latency; cell-free gains of 4–6 $\times$ in energy efficiency vs. cellular) (Lou et al., 2023, Svistunov et al., 22 Oct 2025).
Vertical HetNets (vHetNets): Integrated HAPS–terrestrial networks attain higher coverage and fairness, with careful joint MIMO beamforming and user association suppressing inter-tier interference (e.g., 25% min-SE gain, 2 $\times$ edge rate improvement in urban deployments) (Shamsabadi et al., 2023, Abbasi et al., 2023).
Hybrid SMBS/Relay/RIS Operation: Multi-mode HAPS dynamically switches operational paradigms (computing, relay, RIS) by maximizing energy efficiency, with closed-form analytic switching rules (Alfattani et al., 2022).

Network slicing, MC (Multi-Connectivity), and AI-based resource control are key for multi-service slicing (eMBB, URLLC, mMTC) in such synthesized architectures (Svistunov et al., 22 Oct 2025, Kurt et al., 2020).

5. Optimization, Resource Allocation, and Channel Models

Advanced resource management in HAPS-based networks leverages:

PHY-Layer Innovations: mMIMO, NOMA (non-orthogonal multiple access), power-domain multiplexing, and massive beamforming—subject to energy, coverage, and user-QoS constraints (Javed et al., 2024, Javed et al., 2022).
AI-Driven Control: Reinforcement learning and federated approaches optimize HAPS positioning, beam selection, and handover for maximizing sum-rate, minimizing energy or handover events (Svistunov et al., 22 Oct 2025, Kurt et al., 2020).
Channel Models: Predominantly Rician fading (K-factor 10–20 dB) in LoS, with log-normal or Nakagami-m components for NLoS/obstacle-rich regions; atmospheric fading dominates FSO links but can be mitigated with adaptive optics and beam steering (Lou et al., 2023, Elkhazraji et al., 8 Nov 2025).
Link Budget Formulation: Combined access/feeder path loss, antenna gains, and noise expressions determine achievable per-user and aggregate rates (Xing et al., 2021, Kement et al., 2022).
Resilience and Survivability: Survivable FSO mesh architecture is validated with dual uplink/backup topology, redundancy boosting end-to-end availability $>97\%$ with only marginal equipment cost increase (Truong et al., 2022). RF-based HAPS can wirelessly harvest energy via other HAPS nodes, enabling continuous operation even during solar obscuration (Tuylu et al., 6 Jan 2026).

6. Application Domains and Field Validations

HAPS systems are deployed or prototyped in diverse domains:

Urban and Rural Broadband: Rapid NA/NB-IoT and eMBB extension, massive device support (up to millions per cell) with low propagation delay ( $<$ 0.1 ms/km) (Svistunov et al., 22 Oct 2025, Abbasi et al., 2023).
Disaster/Emergency Recovery: Rapid restoration (e.g., restoring 70% UMTS throughput, as demonstrated by Loon), real-time imaging, and edge computing for incident response (Jaafar et al., 2021, Svistunov et al., 22 Oct 2025).
Intelligent Transportation Systems (ITS): HAPS-ITS nodes provide CAV command/control, high-throughput vehicle links (up to 450 Mbps/vehicle for Level 5 autonomy), multi-Tbps backhaul, and sub-ms response (Jaafar et al., 2021).
ISAC (Integrated Sensing and Communication): Co-located FSO/DIAL sensor payloads enable continuous, high-precision atmospheric gas monitoring ( $<$ 0.1 ppm), disaster gas-leak detection (ppb sensitivity), and environmental mapping (Elkhazraji et al., 8 Nov 2025).
Green Data Center Extension: Stratospheric flying data centers with up to 40-server payloads, ∼12–14% lower energy requirement vs. ground sites, provisioned via modular, open-air cooling racks (Abderrahim et al., 2023, Abderrahim et al., 2023).

Field trials validate multi-week platform endurance (Airbus Zephyr, Loon), Gbps-class downlink, and QKD via FSO from HAPS altitudes, confirming system-level feasibility (Chu et al., 2020, Svistunov et al., 22 Oct 2025).

7. Future Directions, Challenges, and Standardization

Research is advancing on several open fronts:

Channel and Interference Modeling: Non-stationary, multi-environment fading, high-mobility handover, and accurate RIS-aided propagation models.
Standardization and Regulation: Harmonization of HAPS frequency bands; 3GPP NTN integration for seamless HAPS–terrestrial–satellite handover; aviation regulation compliance (Svistunov et al., 22 Oct 2025, Elkhazraji et al., 8 Nov 2025).
AI-Native Network Management: Safe, multi-agent reinforcement learning for fleet deployment, resource slicing, station-keeping, and resilience (Kurt et al., 2020, Svistunov et al., 22 Oct 2025).
Energy Sustainability: Advanced solar/fuel-cell technologies, cooperative HAPS-HAPS energy sharing, and joint trajectory/comms optimization for multi-month self-sustained missions (Tuylu et al., 6 Jan 2026, Javed et al., 2022).
Physical Layer Security: RIS-enabled secrecy, blockchain-based access controls, and quantum key distribution relays from HAPS (Alfattani et al., 2022, Chu et al., 2020).
Integration with ISAC, Multi-Tier NTN, and Edge Services: Coherent design of sensing, communication, and AI/edge compute payloads; mesh inter-HAPS FSO backbones and LEO satellite relays (Elkhazraji et al., 8 Nov 2025, Lou et al., 2023).

Practical deployment will require solutions to payload integration limits, energy management under variable insolation, and FSO/radio mesh reconfiguration for persistent coverage. Standardized APIs and protocols for HAPS operation, mode switching, and spectrum coexistence will be crucial for large-scale adoption (Alfattani et al., 2022, Svistunov et al., 22 Oct 2025).

For a comprehensive technical treatment, see (Kement et al., 2022, Abderrahim et al., 2023, Lou et al., 2023, Elkhazraji et al., 8 Nov 2025, Alfattani et al., 2022, Javed et al., 2022, Song et al., 2023, Jaafar et al., 2021, Abderrahim et al., 2023, Javed et al., 2024, Svistunov et al., 22 Oct 2025, Chu et al., 2020), and (Xing et al., 2021).