6G Non-Terrestrial Networks and the Road Ahead

Integrating Space and Ground

6G envisions a unified air interface spanning terrestrial cellular, LEO satellite, aerial (UAV), and even sub-orbital platforms — all using a common waveform and protocol stack. OTFS is the natural candidate: it handles arbitrary Doppler (LEO, aerial, terrestrial vehicles) and is compatible with the cell- free architecture already standardized for terrestrial 6G. This section maps out how LEO-OTFS integrates with broader 6G Non- Terrestrial Networks (NTN) and what the road to deployment looks like.

Definition:
6G Non-Terrestrial Networks (NTN)

6G Non-Terrestrial Networks (NTN) integrate:

Terrestrial (ground BSs): mature cellular infrastructure.
LEO (500-2000 km): high throughput, low latency.
MEO (2000-20000 km): balanced rate/coverage (not currently deployed at scale).
GEO (35{,}786 km): global coverage, high latency. Legacy.
HAPS (high-altitude platform stations, ~20 km): experimental stratospheric balloons/UAVs.
UAVs (below ~10 km): drones as relays or users.

Unified protocol stack: all tiers share:

Physical layer: OTFS modulation with adaptive numerology.
Cell-free coordination across all tiers via ground CPU/RAN.
Handover: soft handover within tier, managed hard handover between tiers.

Target: single UE device can switch seamlessly between terrestrial, LEO, HAPS, UAV depending on location, mobility, service type.

Theorem: OTFS Compatibility Across NTN Tiers

OTFS handles Doppler across the full NTN mobility range:

Terrestrial vehicle: $\nu \leq 20$ kHz at 28 GHz.
Commercial aircraft: $\nu \leq 50$ kHz at 28 GHz.
HAPS: $\nu < 10$ kHz (quasi-stationary).
LEO: $\nu \leq 690$ kHz at 28 GHz.
GEO: $\nu < 1$ kHz (essentially stationary from Earth POV).

All values accommodated by OTFS frame size: $N$ Doppler bins, frame duration $T$ , $\Delta\nu = 1/T$ . For $T = 5$ ms: $\Delta\nu = 200$ Hz. For 690 kHz Doppler spread: $N = 3450$ bins required — feasible.

Single waveform from ground to orbit. No modulation switch during UE handover across tiers. Key operational simplification for 6G.

The entire NTN stack — from stationary UE receiving GEO to fighter jet receiving LEO — sees dramatically different Doppler. Classical design required tier-specific waveforms. OTFS, with its DD-domain processing, handles all tiers uniformly. The UE's OTFS receiver does not care whether it's connected to a cell tower, a UAV, a LEO satellite, or GEO — the waveform is the same; only the DD-channel parameters differ.

Proof

Frame parameter range

$N = \max(\nu) \cdot T \cdot \text{safety margin}$ . For all NTN tiers, $N \leq 5000$ feasible.

Decoder universality

MP detection works for any DD-sparse channel. Number of paths: 1-3 (LEO), 5-15 (terrestrial).

Common operation

One physical-layer design, one receiver architecture, one control protocol. $\blacksquare$

Definition:
3GPP NTN Roadmap

3GPP NTN standardization timeline:

Release 17 (2022): Initial NTN support. OFDM-based. Mobile Satellite Services (MSS). Direct-to-smartphone from GEO.
Release 18 (2024): Enhanced NTN. Better Doppler handling via pre-compensation. Still OFDM.
Release 19 (2025-2026): LEO-specific optimizations. Cell reconfiguration. Still OFDM baseline.
Release 20 (2026-2028): 6G convergence. Study item: OTFS for NTN. Evaluating performance, standardization.
Release 21 (2028-2030): 6G foundation. OTFS as candidate for LEO and high-mobility scenarios.
Release 22+ (2030+): Full 6G rollout. OTFS is standard.

Example: A UE's Journey Across 6G NTN

A UE in a car drives from city center to a remote mountain area. Describe the OTFS-NTN connectivity journey.

Solution

City: Terrestrial

Urban 5G/6G cellular. Terrestrial OTFS cell-free. $L = 100$ APs per km². Peak rate: 1 Gbps.

Suburb

Sparser terrestrial APs. LEO fallback available. UE uses terrestrial + LEO simultaneously via cell-free combining. Rate: 500-1000 Mbps.

Highway

Terrestrial APs getting sparse. LEO-OTFS becomes primary. $S = 4$ - $5$ satellites. Rate: 50-100 Mbps. Full reliability.

Mountain area

No terrestrial coverage. Pure LEO-OTFS. $S = 3$ - $4$ satellites. Rate: 30-50 Mbps. 99.99% availability.

Remote cabin

Line-of-sight LEO. $S = 5$ - $6$ satellites. Rate: 100 Mbps. Continuous service — the first time remote users get urban- grade connectivity.

Seamless

Throughout the journey, UE connects via OTFS — the modulation does not change. UE decoder handles arbitrary Doppler and AP configuration automatically.

Theorem: Universal 6G Coverage via LEO-OTFS

For a 6G NTN deployment with $\sim 10{,}000$ LEO satellites (6G- era constellation) and full cell-free OTFS coordination, every point on Earth's surface has:

Minimum simultaneous visibility: $S \geq 3$ satellites (except polar regions, $S \geq 2$ ).
Average rate: $\geq 50$ Mbps for mid-latitude UEs.
Reliability: $\geq 99.99\%$ availability.
Latency: $\leq 10$ ms.

Total 6G NTN capacity: $\sim 100$ Tbps aggregate, serving ~1 billion devices globally.

Consequence: OTFS-based 6G NTN provides the technical foundation for universal broadband access — eliminating the digital divide. Every person on Earth can have mid-range internet connectivity, regardless of location.

This theorem quantifies the transformative impact of 6G NTN. Terrestrial networks cover $\sim 30\%$ of Earth's surface (where people live in dense clusters). The remaining $70\%$ — oceans, deserts, polar regions, mountains — has essentially no connectivity today. 6G LEO-OTFS changes this: global coverage at meaningful rates.

Proof

Constellation coverage

10{,}000 LEO satellites at 550 km altitude provide visibility fraction $\sim 2000/R_\oplus = 0.3$ at any point (per satellite). For $S \geq 3$ : need $n = 10000 \cdot 0.3 = 3000$ satellites visible on average. Achievable with constellation.

Rate bound

Per-sat $\sim 30$ Mbps to ground UE. Macro-diversity $S = 4$ : aggregate $\sim 100$ Mbps. Per-UE scheduling + user-centric clustering: 50 Mbps average.

Reliability

$S \geq 3$ satellites, each 95% availability: aggregate $1 - (0.05)^3 = 99.99\%$ .

Aggregate

Total capacity: 10{,}000 sat × 10 Gbps = 100 Tbps. Users: 1 billion at 50 Mbps ≈ 50 Tbps. Constellation sized for 2 $\times$ capacity margin. $\blacksquare$

Key Takeaway

LEO-OTFS enables universal 6G broadband. 10{,}000 satellites + cell-free OTFS + multi-tier integration → 99.99% global coverage at 50+ Mbps, with 10 ms latency. Every person on Earth with broadband access — the 2030s vision of 6G NTN.

6G NTN Coverage vs Constellation Size

Plot global 6G NTN coverage (% of Earth with $S \geq 3$ satellites) as a function of total constellation size. Overlay current (6k) and projected (20k, 40k) constellations.

Parameters

Satellites10000

Min elevation (°)10

LEO Satellite Doppler Sweep at 28 GHz

Animation of a LEO satellite passing overhead at 7.5 km/s. Shows the Doppler frequency sweeping from

+690

kHz (approach) through 0 (zenith) to

-690

kHz (recession). Side-by-side: OFDM struggling with ICI, OTFS's DD-domain representation remaining stable and sparse. The visual case for OTFS in LEO.

🔧Engineering Note

6G NTN Deployment Roadmap

6G NTN with OTFS rollout plan:

2024-2026: 5G NR NTN (Rel. 17/18) — OFDM-based. Commercial service from select GEO operators (Starlink Direct-to-Cell, etc.). OTFS at research/prototype stage.

2026-2028: 3GPP Rel. 20 study item on OTFS for NTN. Performance evaluations. Standardization work begins.

2028-2030: Rel. 21 includes OTFS as candidate for LEO. Commercial OTFS-enabled satellite launches (Starlink Gen 3, Kuiper, OneWeb Next).

2030-2035: Rel. 22+ full 6G deployment. OTFS standard for high-mobility NTN. Seamless terrestrial-LEO integration. 10+ billion devices served globally.

Beyond 2035: Expected convergence — 6G NTN becomes the dominant global broadband fabric. Terrestrial cellular serves as "local high-capacity" supplement. LEO-OTFS is the baseline.

Practical Constraints

•
2024-2026: OFDM NTN (legacy)
•
2026-2028: OTFS NTN study item (Rel. 20)
•
2028-2030: OTFS NTN standardization (Rel. 21+)
•
2030+: Full 6G deployment with OTFS baseline
•
2035+: 10B+ devices on global 6G NTN

Common Mistake: NTN Spectrum Coordination Is Complex

Mistake:

Assuming NTN spectrum can be freely allocated. Terrestrial and satellite systems share many bands (Ka, Ku, V), requiring careful coordination to avoid interference.

Correction:

ITU and national regulators coordinate NTN/terrestrial spectrum allocations. Typical scheme:

Terrestrial: primary use below ~40 GHz for cellular.
NTN: primary use above ~12 GHz for downlink, some uplink.
Shared bands: tight interference budgets. NTN must meet terrestrial PFD (power flux density) limits at Earth's surface.

OTFS-NTN adoption may trigger re-coordination as per-user throughput increases. Expect regulatory friction; 10+ year process for new bands. Operational OTFS-NTN at current bands: feasible with existing 3GPP NTN specs (updated for OTFS in Rel. 21+).

Why This Matters: From Chapter 1 to Chapter 18: The OTFS Arc

This chapter closes the main arc of the OTFS book. Chapter 1 established the DD-domain as the natural signal space for high-mobility channels. Chapters 2-5 developed the mathematical foundations (Zak transform, symplectic Fourier transform). Chapters 6-10 derived OTFS modulation and detection. Chapters 11-15 applied OTFS to ISAC (integrated sensing and communication) and automotive V2X. Chapters 16-18 extended to network scales: MIMO-OTFS, cell-free, LEO satellite. The remaining chapters (19-22) will survey 6G standardization, pulse shaping, machine learning for OTFS, and open problems. The thread throughout: OTFS is the correct signal space for high-mobility, delay- dispersive, multi-transceiver wireless — a 6G backbone.

OTFS vs OFDM Performance at Orbital Velocity Chapter Summary