⚡ electrical engineering

A Disaster-Aware Integrated TN-NTN System-Level Simulator for Resilient 6G Wireless Networks

This paper presents a lightweight, 3GPP-compliant system-level simulator designed to evaluate the resilience and performance trade-offs of integrated terrestrial and non-terrestrial 6G networks under partial-failure disaster scenarios, demonstrating how hybrid operations can maintain service reliability through effective traffic migration.

Original authors: Donglin Wang, Anjie Qiu, Qiuheng Zhou, Hans D. Schotten

Published 2026-05-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Donglin Wang, Anjie Qiu, Qiuheng Zhou, Hans D. Schotten

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a city's mobile network as a busy highway system. Under normal conditions, everyone drives on the Terrestrial Network (TN)—the local roads, tunnels, and bridges built right into the ground. These roads are fast, efficient, and perfect for daily commutes.

However, when a disaster strikes (like a massive earthquake or storm), these local roads can get blocked, damaged, or completely destroyed. If the local roads are gone, traffic stops, and people are stranded.

This paper introduces a digital "crash-test simulator" designed to figure out how to keep traffic moving when the local roads fail. It tests a backup plan using Non-Terrestrial Networks (NTN)—think of these as a fleet of delivery drones or satellites flying high above the city. They can't be as fast as the local roads, but they can't be destroyed by a storm on the ground.

Here is how the paper breaks down this concept in simple terms:

1. The Problem: When the Ground Fails

The authors note that our modern world relies heavily on cell towers. If a disaster knocks out power or destroys the towers, the network collapses. While satellites (LEO) and drones (UAVs) exist to help, we didn't have a simple, fast way to test exactly how well they work together with the ground network during a crisis. Existing tools were either too complicated (like simulating every single radio wave) or only looked at one type of network in isolation.

2. The Solution: A "What-If" Simulator

The team built a lightweight simulator. Think of this not as a video game, but as a sophisticated spreadsheet that runs thousands of "what-if" scenarios in seconds.

The Setup: They created a virtual city with 10 cell towers (the ground network) and a ring of satellites overhead.
The Disaster: They randomly "break" some of the cell towers (simulating a disaster where 50% might go down).
The Rescue: The simulator automatically reroutes the people who lost their local connection to the satellites. It also checks if the remaining ground towers are too crowded and moves some people to the satellites to prevent gridlock.

3. The Rules of the Game

The simulator follows specific rules based on real-world standards (3GPP):

The "Panic" Factor: When the disaster hits, people don't just stay put; they move around (simulated as "panic mobility").
The Handover: If your local tower is broken, you are forced to switch to the satellite. If your local tower is still working but super crowded, the system might voluntarily move you to the satellite to keep things moving.
The Bottleneck: The satellites have a "feeder link"—a giant pipe connecting them to the internet core. If too many people try to use the satellite at once, this pipe gets clogged, slowing everyone down.

4. What the Simulator Found

The researchers ran the simulation with different numbers of people (from 100 to 500 users) and different levels of disaster severity. Here is what they discovered:

Normal Times (No Disaster): The ground network is the clear winner. It's faster and has lower "latency" (delay). The satellite is slower because the signal has to travel all the way to space and back.
Disaster Times: When the ground network is damaged, the hybrid system (Ground + Satellite) saves the day.
- Reliability: Even if half the towers are gone, the system keeps working. The "Packet Reception Ratio" (how many messages get through) actually goes up because the satellite ensures no one is left completely offline.
- The Trade-off: You get reliability, but you pay for it with speed. As more people rely on the satellite, the average speed drops, and the delay increases. It's like switching from a sports car to a bus: the bus gets you there when the road is blocked, but it's slower and less comfortable.
The "Feeder" Limit: They found that if the satellite's connection pipe (feeder) is too small, it becomes a bottleneck. Increasing the pipe size helps a lot up to a point (around 450 Mbps), but after that, the speed is limited by how many people are trying to use the satellite, not the pipe size itself.

5. The Big Picture

The paper concludes that this simulator is a useful tool for engineers. It proves that while the ground network is best for speed, and the satellite network is best for coverage, combining them creates a resilient system.

When a disaster hits, the system automatically shifts traffic to the sky. It's not perfect (it's slower), but it ensures that communication doesn't stop entirely. The simulator helps planners figure out exactly how many satellites they need and how big their "pipes" should be to handle a crisis without getting overwhelmed.

In short: The paper presents a digital test lab that shows how mixing ground towers with sky-based satellites can save the day when a disaster knocks out the ground, ensuring we stay connected even when the roads are broken.

Technical Summary: A Disaster-Aware Integrated TN-NTN System-Level Simulator for Resilient 6G Wireless Networks

Problem Statement
Modern societies rely heavily on terrestrial cellular networks (TN), which are inherently vulnerable to infrastructure damage, power outages, and congestion during natural disasters and large-scale failures. While the 3rd Generation Partnership Project (3GPP) has standardized Non-Terrestrial Networks (NTN)—including Low Earth Orbit (LEO) satellites, High-Altitude Platform Stations (HAPS), and Unmanned Aerial Vehicles (UAVs)—as a resilience mechanism for 5G-Advanced and 6G, evaluating the joint behavior of integrated TN-NTN systems under disaster conditions remains challenging. Existing literature often focuses on high-fidelity physical-layer modeling or isolated performance of either TN or NTN, lacking dedicated tools that capture the coupling, service migration, and recovery dynamics required for system-level resilience analysis in chaotic post-disaster environments.

Methodology
The authors present a lightweight, parameter-driven system-level simulator designed to evaluate post-failure fallback behavior in integrated TN-NTN networks. The framework adheres to 3GPP Release 17/18 modeling principles and utilizes a static post-failure snapshot approach rather than a full time-evolving recovery process.

Architecture: The simulation models a heterogeneous network where a terrestrial segment (10 gNBs in a perturbed hexagonal layout) operates alongside an NTN overlay (LEO satellites at 600 km altitude). The system supports dynamic service migration where traffic is offloaded from degraded terrestrial links to satellite links.
Disaster Model: The disaster scenario is defined by a probabilistic failure model where terrestrial gNBs are independently marked as "failed" based on a failure probability ( $p_f$ ).
Operational Logic: The simulator employs a two-stage partitioning algorithm:
1. Forced Handover: Users attached to failed gNBs are forcibly migrated to the NTN.
2. Proactive Offloading: Based on a configured aggressiveness parameter ( $\eta$ ), additional users on surviving but overloaded or weak terrestrial cells are proactively switched to NTN.
3. Association Metric: User association is determined by a load-aware score ( $\Gamma_{u,b}$ ) that penalizes heavily loaded cells, calculated as the difference between SINR and a load penalty term.
Performance Modeling: The simulator calculates Key Performance Indicators (KPIs) including throughput, Packet Reception Ratio (PRR), and latency. It incorporates simplified overhead models, including fixed migration penalties, reassociation penalties for terrestrial users, and a shared feeder-capacity constraint for the NTN.
Implementation: The tool is implemented as a Python framework using Monte Carlo simulations to capture stochastic geometry, failure sampling, and shadowing effects.

Key Contributions

Lightweight Simulation Engine: The paper introduces a tractable system-level simulator that bridges the gap between high-fidelity specialized tools and isolated network studies, specifically tailored for "what-if" experimentation in disaster scenarios.
Integrated Failure Modeling: The framework explicitly models the coupling between terrestrial and satellite links, capturing the dynamics of service migration, congestion relief, and the trade-offs between terrestrial efficiency and non-terrestrial resilience.
Quantitative Trade-off Analysis: The simulator provides a mechanism to analyze the capacity-delay tradeoff of terrestrial operations against the reliability and stability of non-terrestrial services under varying user loads, disaster severities, and NTN provisioning levels.

Results
The simulator was evaluated across user loads ranging from 100 to 500 UEs and varying gNB failure probabilities ( $p_f$ ). Key findings include:

Capacity-Delay Tradeoff: Under nominal conditions, terrestrial networks offer the lowest latency but suffer throughput degradation as load increases. NTN provides stable latency (dominated by propagation delay) but lower per-user capacity.
Hybrid Resilience: In disaster mode (e.g., $p_f = 0.5$ ), the hybrid TN-NTN mode offers an intermediate tradeoff. While throughput is lower than nominal TN due to migration penalties and feeder constraints, the system maintains service continuity.
Disaster Severity Impact: As gNB failure probability increases, the realized NTN fallback ratio ( $\hat{\eta}$ ) increases monotonically. This leads to a slight initial throughput gain (due to congestion relief) followed by a sharp decline as the system becomes fully dependent on the feeder-limited NTN path. However, the Packet Reception Ratio (PRR) improves significantly (from 0.67 to 0.94) as NTN fallback ensures connectivity for users who would otherwise be disconnected.
Feeder Bottleneck: Sensitivity analysis reveals that NTN feeder capacity is a dominant bottleneck only below a moderate threshold (approx. 450 Mbps). Beyond this, performance is limited by access-side sharing and migration overhead rather than feeder scaling.

Significance and Claims
The paper claims that the proposed framework serves as a "tractable tool" for studying wireless network resilience and traffic management in future integrated 6G systems. The authors position the simulator not as a calibrated protocol-stack implementation, but as an exploratory engine for rapid comparative analysis.

The significance lies in its ability to identify critical core bottlenecks and congestion factors that high-fidelity specialized tools might overlook due to computational complexity. The results demonstrate that while terrestrial networks remain advantageous for low-delay services under moderate loads, NTN provides essential continuity when terrestrial access degrades. The hybrid operation is presented as a practical solution offering a balanced tradeoff between capacity, reliability, and delay. The authors conclude that this tool is suitable for "what-if" analysis prior to moving to higher-fidelity tools or external validation benchmarks, and they note that future work will extend this static abstraction to include time-varying satellite dynamics and richer failure-recovery processes.