5G Non-Terrestrial Networks (NTN) Overview

• 5G NTN is a network architecture that uses satellites and airborne platforms to extend 5G service beyond terrestrial networks.
• A multi-layer design integrates spaceborne segments (LEO/MEO/GEO), ground gateways, and relay nodes to optimize latency, coverage, and capacity.
• Adapted physical and MAC layers address challenges such as Doppler shifts, high path loss, and spectrum sharing to support versatile applications.

A 5G Non-Terrestrial Network (NTN) is defined as a 5G-compatible radio access network utilizing spaceborne or airborne platforms—such as satellites in Low Earth Orbit (LEO), Medium Earth Orbit (MEO), Geostationary Earth Orbit (GEO), High Altitude Platforms (HAPS), or Unmanned Aerial Vehicles (UAVs)—to extend coverage above the Earth’s surface and interoperate seamlessly with terrestrial 5G New Radio (NR) infrastructure. NTNs are integral to the 3GPP standards (notably Release 17 and beyond) and support ubiquitous connectivity, capacity augmentation, and robust global backhaul, particularly in areas beyond terrestrial coverage or in high-density environments (Azari et al., 2021).

1. System Architectures and Key Components

The architecture of 5G NTNs is organized into multi-layer, multi-band frameworks integrating space (LEO/MEO/GEO satellites), air (HAPS/UAVs), and terrestrial segments. Core components include:

• Satellite Segments: GEO satellites (~35,786 km) provide static, wide-area coverage for broadcast and backhaul with inherent high round-trip delays (~500 ms RTT). MEO satellites balance coverage and latency (~100 ms RTT). LEO satellites (~500–2,000 km) offer low-latency (~5–25 ms RTT), high Doppler, and frequent handovers via moving spot-beams.
• Ground Segment: Gateways (GW) interface the satellite feeder links to the 5G Core (5GC), with functionality physically split between central units (CUs) and distributed units (DUs), depending on the payload architecture (transparent vs. regenerative) (Lin et al., 2021, Guidotti et al., 2022).
• Relay and Integration Nodes: On transparent payloads, satellites act as bent-pipe relays. Regenerative payloads, expected to dominate future 5G-A/6G, integrate part of the gNB CU/DU onboard, enabling in-space processing, scheduler flexibility, and inter-satellite routing via ISLs.
• User Equipment (UE): 5G NR UEs can be standard handhelds, VSATs, IoT modules, or mobile relay nodes, typically equipped with GNSS receivers for position and timing required in high Doppler and delay environments.

A multi-layered architecture, combining LEO satellite backhaul, passive RIS-augmented terrestrial cells, and advanced mMIMO beamforming, achieves significant improvements in coverage, capacity, reliability, and cost-efficiency, notably in scenarios with thousands of simultaneous UEs (e.g., large events) (Farré et al., 12 Jun 2025).

2. Physical Layer Adaptations and Channel Models

NTN physical layer design requires adaptations to counteract unique impairments:

• Waveform and Numerology: OFDM is retained but subcarrier spacing (SCS, e.g., 15–120 kHz) may be adjusted for Doppler tolerance (LEO: ±50 kHz at 2 GHz), shortening slot durations respectively. Extended cyclic prefix (CP) can absorb large differential delays (Guidotti et al., 2022, Azari et al., 2021).
• Path Loss and Link Budget: The core formula governing received power is

$$P_r = P_t + G_t + G_r - L_{\mathrm{fs}} - L_{\mathrm{atm}} - L_{\mathrm{rain}} - L_{\mathrm{pol}} - L_{\mathrm{misc}}$$

where $L_{\mathrm{fs}} = 20\log_{10}(4\pi f_c d / c)$ (FSPL), with large absolute values in GEO/MEO scenarios (FSPL ≈ 170–190 dB) (Hernandez et al., 2023, Azari et al., 2021).

• Doppler Compensation: UEs estimate Doppler via GNSS and satellite ephemeris or, if GNSS is unavailable, by multi-reference-signal estimation using OFDM pilot tones spaced across the band, achieving practical compensation with ~1 kHz error in LEO (Lin et al., 2021).
• Shadowing, Rain, and Fading: Channels modeled per 3GPP TR 38.811/38.821, with losses from atmospheric absorption, rain fade (especially at Ka-band), and site-specific shadowing. Adaptive coding/modulation (ACM), site diversity, and phased-array beamforming are standard mitigations.

RIS (Reconfigurable Intelligent Surface) panels, passive devices co-located near gNodeBs, can close coverage holes, extend cell radii (e.g., 45→70 m in stadium deployments), and provide >10 dB nominal reflection gain with low power consumption (<10 W) (Farré et al., 12 Jun 2025).

3. MAC, RRM, and Network Layer Enhancements

• Timing Advance and Synchronization: UEs require both a “common TA” (satellite-to-gNB RTT) and a UE-computed TA (satellite-to-UE RTT, GNSS-assisted) to align their uplink timing, especially under large satellite-induced delays (Lin et al., 2021).
• HARQ and Scheduling: Typical terrestrial NR supports up to 16 HARQ processes. In GEO NTNs, disabled HARQ feedback and greater process recycling (e.g., up to 32 HARQ processes) are used. RLC and PDCP retransmission handle any reliability compensation (Lin et al., 2021, Hou et al., 25 Sep 2025).
• Mobility Management and Handover: NTN handover enhancements include conditional or predictive handover based on UE location and satellite movement, group handover (GHO) to batch thousands of UEs, reducing signaling storms and latency by ~88% compared to baseline per-UE HO (Zhang et al., 20 Mar 2024).
• Multi-Connectivity (MC): Dual connectivity, bearer duplication, and dynamic Xn interface management across multiple satellites and terrestrial/gNBs are supported. Packet duplication across independent links improves reliability (single-link $P_{\mathrm{succ}} ≈ 99\%$, MC $P_{\mathrm{PD}} ≈ 99.99\%$), and resource-aware scheduling maximizes end-to-end efficiency (Majamaa, 2023).

4. Resource Management, Spectrum Sharing, and Interference

NTNs introduce novel resource optimization challenges due to the shared use of S-, C-, and Ka-bands and the coexistence with terrestrial networks.

• Joint Optimization: Sum-rate maximization across UEs is achieved by jointly allocating spectrum (S-, C-, Ka-), beamforming (mMIMO), and RIS phase shifts, often via alternating optimization with water-filling algorithms. Real-time adaptability is feasible with convergence within 10 iterations (Farré et al., 12 Jun 2025).
• Dynamic Spectrum Sharing (DSS): Centralized spectrum management servers (SMS) dynamically partition Resource Blocks between TN and NTN based on instantaneous load, with interference thresholds protecting TN QoS and enabling opportunistic reuse (~10–15% spectral efficiency gains); guard bands and coordinated frequency group assignments prevent excessive cross-system interference (Martikainen et al., 2023).
• Cell-Free MIMO: Federated MIMO swarms of cooperating LEO satellites, using location-only beamforming (LB-MMSE), enable terabit-scale capacity with robust performance despite CSI aging and realistic localization errors. Swarm-level normalization algorithms (sSPC, sMPC) guarantee per-satellite power constraints (Guidotti et al., 2023).

5. End-to-End Services, QoS, and Application Performance

NTNs underpin a rich set of services and use cases across the mobility, IoT, and broadband spectrum:

• Coverage Extension: NTNs bridge digital divides in remote, rural, maritime, and aeronautical scenarios, using LEO for low-latency, seamless handover; GEO for continent-scale coverage and multicast/broadcast offload (Guidotti et al., 2022, Azari et al., 2021).
• IoT Mass Connectivity: NB-IoT/eMTC support over NTNs for sensor telemetry and low-rate messaging in remote areas with discontinuous coverage, using extended PRACH formats and low-duty-cycle access (Azari et al., 2021).
• Quality of Experience: Compared to DVB-S2/RCS2, 5G NTN offers lower jitter (4 ms vs. 13 ms), faster video TTFF (56 s vs. 114 s), and higher file-transfer throughput (2.2×), especially for interactive and delay-sensitive apps. DVB remains optimal for high-throughput broadcast/bulk downloads in stable SNR regimes (Kumar et al., 30 Sep 2025, Sormunen et al., 19 Feb 2025).
• Capacity Metrics: Downlink capacity for LEO/S-band handheld is ~600 Mbps per satellite; Ka-band/VSAT terminals achieve ~7 Gbps (Sedin et al., 2020). Area capacity density ranges from ~1–10 kbps/km² (S-band) to ~14–120 kbps/km² (Ka-band).
• Reliability and Availability: Outage rates can be held below 1% even in ultra-dense scenarios via SINR-aware resource optimization and group handover protocols (Farré et al., 12 Jun 2025, Zhang et al., 20 Mar 2024).

6. Security, Orchestration, and AI-Driven Frameworks

Integration of TN and NTN segments increases attack surfaces and operational complexity:

• Security Architecture: Multi-layer frameworks span space, air, ground, and cloud/edge. Key threats include jamming, spoofing, DDoS, supply-chain tampering, and discontinuous satellite coverage loss (Maric et al., 7 Aug 2025).
• AI-Native Cloud Security: Systems deploy per-slice zero-trust credentials, real-time anomaly detection (e.g., LSTM autoencoders, Transformer classifiers), federated learning, temporal-logic policy enforcement, and automated orchestration. Resilience metrics (fraction of service maintained, MTTR) quantify robustness (Maric et al., 7 Aug 2025).
• AI-Open-RAN Integration: AI modules for KPI (e.g., SINR) forecasting, mobility learning, Doppler and blockage prediction, and delay-aware scheduling are embedded at all layers. LSTM-based forecasting of SINR under mobility can reduce outage time by ~30% in testbed emulation (Do, 14 Nov 2025).
• Slicing & Virtualization: Resource segmentation into slices with dynamic reconfiguration and guaranteed isolation (Cov(U_i, U_j)=0) for throughput, latency, and security. Cloud-native orchestration and real-time monitoring underpin scalable, resilient NTN deployments (Maric et al., 7 Aug 2025).

7. Positioning, Synchronization, and Protocol Evolution

NTN-enhanced positioning is a major paper item in 5G Release 18 and beyond:

• LEO-Based Localization: Single-LEO round-trip time (RTT) methods achieve coarse verification (<10 km accuracy). Multi-LEO TDOA+RTT architectures realize sub-meter positioning (<1 m PEB), and augmentations with minimal GNSS (2 satellites) approach similar accuracy (Dureppagari et al., 2023).
• Performance Bounds: CRLB analysis for N RTT or M TDOA measurements provides position error bounds:

$$\mathrm{CRLB}_{\mathrm{single}} = (\sum_{i=1}^N \frac{1}{\sigma_r^2} u_i u_i^T)^{-1}$$

$$\mathrm{CRLB}_{\mathrm{multi}} = (\mathbf{G}^T \mathbf{\Sigma}^{-1} \mathbf{G})^{-1}$$

• Technical Gaps: NTNs face non-standardized signal designs for positioning, PRS timing, rapid ephemeris update requirements, per-satellite frequency/time compensation for multi-LEO TDOA, and integration challenges at the LMF, protocol, and UE modem levels. Enhanced uplink (multi-RTT), Doppler/FDOA methods, and robust multi-orbit satellite coordination are open research areas (Dureppagari et al., 2023).

8. Standardization, Field Trials, and Future Directions

• 3GPP Roadmap: NTN integration is established in Release 17 (transparent payloads), matured in Release 18 (enhanced NR_NTN, IoT_NTN), and is projected to include regenerative payloads, advanced beam management, asynchronous MC/CA, RIS integration, and AI-driven RRM in Rel. 19/20 and future 6G (Guidotti et al., 2022).
• Industrial Trials: Ongoing deployments and field trials span GEO NB-IoT (Intelsat), LEO integration (Viasat, OneWeb, Starlink), and multi-segment demonstrations (5G-VINNI, 5GENESIS, SATis5), with anticipated hybrid commercial service by major operators (Orange, Vodafone) (Hernandez et al., 2023, Azari et al., 2021).
• Research Challenges: The path to 6G NTN includes sub-10 cm positioning, Tbps-class in-space routing, quantum-resilient security, optical feeder links, ultra-tight synchronization, and sustainable space–air platforms. AI/ML methods for RRM, dynamic spectrum sharing, and cross-layer optimizations are central to this evolution (Guidotti et al., 2022).

---

The 5G NTN landscape encompasses a harmonized integration of spaceborne and terrestrial assets through advances in architecture, waveforms, MAC/RRM algorithms, spectrum sharing, security, and AI-driven orchestration, aiming to deliver global, resilient, high-capacity, and flexible networks for diverse use cases ranging from IoT to mission-critical services, and forming the backbone for the upcoming 6G era (Azari et al., 2021, Farré et al., 12 Jun 2025, Guidotti et al., 2022).