Bridging the Connectivity Gap: The Technical Evolution of 5G Non-Terrestrial Networks

The long-standing challenge of universal wireless coverage is nearing a solution as engineers work to integrate satellite constellations directly into the 5G ecosystem. By moving beyond ground-based towers to a unified Non-Terrestrial Network (NTN), the telecommunications industry aims to eliminate "dead zones" in even the most remote regions of the planet. A comprehensive study titled "5G NR Non-Terrestrial Networks: Open Challenges for Full-Stack Protocol Design," authored by a team of researchers from the University of Padova and Toyota Motor North America, outlines the critical architectural shifts required to make this vision a reality. As 5th generation (5G) networks continue to evolve, the research highlights that the integration of Terrestrial Networks (TNs) and NTNs represents the next major frontier in global connectivity, promising a seamless transition between cellular towers and orbiting satellites.

The Rise of Non-Terrestrial Networks (NTN)

The concept of Non-Terrestrial Networks encompasses a diverse array of platforms situated above the Earth’s surface, including Low Earth Orbit (LEO) satellites, High Altitude Platforms (HAPs) such as balloons or airships in the stratosphere, and Unmanned Aerial Vehicles (UAVs). While traditional satellite internet has existed for decades, it was largely a proprietary and isolated technology. The current evolution, led by the 3rd Generation Partnership Project (3GPP), seeks to standardize these platforms within the 5G New Radio (NR) framework. This integration is designed to serve three primary objectives: expanding wireless coverage to underserved rural and remote areas, providing resilient backup for emergency communications during natural disasters, and offloading traffic from highly congested urban environments where terrestrial infrastructure is pushed to its limits.

According to the research team, led by Francesco Rossato, Mattia Figaro, and Alessandro Traspadini of the University of Padova, LEO satellites are particularly attractive for these networks. Operating at altitudes between 500 and 2,000 kilometers, LEO constellations offer significantly lower latency than traditional Geostationary (GEO) satellites, which sit much higher at 35,786 kilometers. This proximity allows for the delivery of broadband-speed internet and supports a wide range of modern applications, from autonomous vehicle navigation to environmental monitoring and smart grid management. The study emphasizes that while LEO satellites provide wide-area connectivity, their rapid orbital movement introduces a unique set of technical complexities that terrestrial 5G was never originally designed to handle.

Technical Hurdles in Full-Stack Protocol Design

The transition from ground-based towers to space-borne base stations necessitates a fundamental rethink of the "full-stack" protocol design, spanning the Physical (PHY) layer to the Transport layer. One of the most significant hurdles identified by the researchers is path loss. Signals traveling through the atmosphere are subject to severe attenuation caused by rain, clouds, and ionospheric scintillation. Furthermore, because satellites move at immense speeds relative to users on the ground, they generate significant Doppler shifts—frequency changes that can disrupt the synchronization between the device and the network. If these shifts are not precisely compensated for, the connection becomes unstable or fails entirely.

At the Medium Access Control (MAC) layer, the primary challenge is the sheer distance. Even at the speed of light, the propagation delay to a satellite is vastly greater than the delay to a nearby cell tower. This delay complicates critical procedures such as channel estimation, resource allocation, and scheduling. For example, traditional 5G networks use a process called Hybrid Automatic Repeat reQuests (HARQ) to handle data errors. In a terrestrial setting, the "wait time" for a retransmission is milliseconds; in a satellite network, this delay can stall the entire data flow. The researchers argue that without substantial modifications to how resources are managed and how handovers between satellites are performed, the network will suffer from massive bottlenecks and reduced throughput.

3GPP Standardization and 5G NR-NTN

The path toward a unified network is being codified through the 3GPP standardization process. The researchers detail the journey from Release 17, which introduced the first specifications for direct satellite-to-phone communication, toward future releases such as Release 20. A key assumption in current standards is the "transparent" payload model, where the satellite acts as a "bentpipe" relay, simply amplifying and forwarding signals between a ground gateway and the user's device. However, as the technology matures, there is a push toward "regenerative" payloads where the satellite itself performs onboard processing, effectively acting as a base station (gNB) in orbit.

To validate their theories, the research team, which includes experts like Michele Zorzi and Marco Giordani from the University of Padova and Takayuki Shimizu from Toyota Motor North America, conducted extensive end-to-end simulations using the ns-3 discrete-event simulator. Unlike much of the existing literature which remains conceptual, this study provided numerical evidence on how specific settings impact network performance. Their simulations demonstrated the critical importance of Guard Periods (GPs) in Time Division Duplexing (TDD) and showed how differential delays in large satellite cells can lead to timing misalignments that degrade the user experience. This empirical approach is vital for the 3GPP as it moves to refine standards for the 5G NR-NTN ecosystem.

Detailed Findings: Retransmissions and Transport Latency

The researchers’ findings in the ns-3 simulation environment revealed that the number of HARQ processes—the mechanisms that manage data retransmissions—must be significantly increased for satellite networks. In standard terrestrial 5G, a few processes are sufficient, but in the NTN context, the long round-trip time (RTT) means that many more processes must run in parallel to keep the data link active. Without this adjustment, the transmitter spends the majority of its time waiting for acknowledgments rather than sending new data. Additionally, the study highlighted a "stalling" effect where the MAC layer’s inability to keep up with the long delays causes the higher-level Transmission Control Protocol (TCP) to drastically reduce its transmission rate, further crippling the connection speed.

The team also investigated the impact of large cell sizes. A single satellite beam can cover hundreds of square kilometers, leading to "differential delay," where users at the center of the beam experience different propagation times than those at the edge. The simulation results suggested that the network must implement sophisticated timing advance mechanisms to ensure that signals from different users do not collide when they arrive at the satellite. These findings underscore the necessity of a "satellite-aware" protocol stack that can dynamically adjust to the orbital dynamics and the vast distances involved in non-terrestrial communication.

Commercial Implications: Starlink and the Move to Standardized 5G

The commercial landscape for NTNs is currently dominated by major satellite providers like SpaceX’s Starlink, which has already begun deploying "direct-to-cell" technology. However, many current satellite services rely on proprietary hardware and software. The research by Rossato et al. suggests a major industry shift: moving from these closed, proprietary systems to standardized 5G hardware. This shift would allow standard consumer smartphones to connect to satellites without requiring specialized antennas or chips, a development that would commoditize satellite connectivity and integrate it into standard cellular plans.

The implications for this are profound, not just for 5G but for the upcoming 6G era. By establishing a standardized NTN foundation now, the industry is setting the stage for a truly 3D network architecture where terrestrial towers, drones, and satellites work in a coordinated mesh. Major automotive players like Toyota, involved in this research, are particularly interested in this for connected vehicle safety. A car moving through a remote mountain pass should, in theory, switch from a lost terrestrial signal to a satellite signal so seamlessly that a high-definition navigation map or an emergency call remains uninterrupted.

What's Next: The Future of Global Connectivity

Looking forward, the research team points to several "open questions" that will define the next decade of telecommunications. Future standardization activities must address more complex routing scenarios, especially for Inter-Satellite Links (ISL), where data is hopped from one satellite to another in space before being sent down to a ground station. Furthermore, the integration of Artificial Intelligence (AI) and Machine Learning (ML) into the protocol stack could allow for predictive handover management, anticipating when a user will lose line-of-sight with a satellite and proactively switching the connection to the next one in the constellation.

The timeline for mainstream adoption is accelerating. With the foundational work laid out in the 3GPP Release 17 and 18, and the rigorous simulation data provided by academic-industry partnerships like the one between the University of Padova and Toyota, the transition from "No Signal" to "Always Connected" is no longer a matter of if, but when. As the paper (currently submitted to IEEE for publication) suggests, the evolution of 5G NR-NTN is not merely an incremental update to cellular technology, but a radical expansion of the boundaries of the internet itself, turning the sky into the next great layer of the global digital infrastructure.

• Primary Authors: Francesco Rossato, Mattia Figaro, Alessandro Traspadini (University of Padova); Takayuki Shimizu (Toyota Motor North America).
• Institutional Support: University of Padova, Italy; Toyota Motor North America Inc., USA; European Union National Recovery and Resilience Plan (NRRP).
• Key Methodology: End-to-end system-level simulations using the ns-3 discrete-event simulator.
• Publication Status: Submitted to IEEE for publication (arXiv:2601.14883v1).