🔢 mathematics

Latency-Constrained Resource Synergization for Mission-Oriented 6G Non-Terrestrial Networks

This paper proposes a latency-constrained resource synergization framework for mission-oriented 6G non-terrestrial networks in post-disaster scenarios, utilizing UAV-mounted edge information hubs to jointly optimize communication and computing resources alongside location planning, thereby achieving a 20% cost reduction while meeting stringent latency requirements.

Original authors: Yueshan Lin, Wei Feng, Yunfei Chen, Yongxu Zhu, Ning Ge, Shi Jin

Published 2026-03-17

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Yueshan Lin, Wei Feng, Yunfei Chen, Yongxu Zhu, Ning Ge, Shi Jin

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 massive earthquake has just hit a remote mountain village. The roads are blocked, cell towers are smashed, and the ground is shaking. The emergency teams need to know exactly what's happening right now to save lives, but they can't get a signal.

This is where the paper comes in. It's about a high-tech rescue plan using drones (UAVs) to act as flying "command centers" in the sky.

Here is the story of the paper, told simply:

1. The Problem: The "Data Flood"

When disaster strikes, hundreds of sensors (like cameras or earthquake detectors) start screaming for help, sending massive amounts of data.

The Bottleneck: If you try to send all that raw data straight to a satellite, the drone's internet connection is too slow. It's like trying to empty a swimming pool with a teaspoon. The data takes too long to upload, and the rescue teams are left waiting in the dark.
The Old Way: Previous solutions just tried to make the "teaspoon" bigger (more bandwidth) or the "pool" smaller (less data), but they didn't think about how to process the water while it was flowing.

2. The Solution: The "Flying Smart Hub"

The authors propose a special drone equipped with an Edge Information Hub (EIH). Think of this drone not just as a mailman, but as a mobile processing factory flying above the disaster zone.

Here is how it works:

The Sensors: The ground sensors send their raw data up to the drone.
The Factory (Edge Computing): Instead of sending everything to space, the drone's onboard computer acts like a smart chef. It tastes the soup (analyzes the data) and throws away the water (discards useless noise), keeping only the delicious broth (the critical information).
The Satellite: The drone then sends just the "broth" up to the satellite. This is much faster and lighter.

3. The Big Challenge: The "Tightrope Walk"

The tricky part is balancing two things:

Computing Power: How strong is the drone's brain? (Too weak, and it can't process fast enough. Too strong, and it's a waste of money and battery.)
Communication Power: How fast is the drone's internet? (Too slow, and the data backs up. Too fast, and you're paying for speed you don't need.)

The paper asks: "What is the perfect recipe?"
If you give the drone a super-fast brain, you might not need such a fast internet connection. If you give it a slow brain, you need a super-fast internet line to keep up. The goal is to find the exact mix that gets the job done in the shortest time for the lowest cost.

4. The "Goldilocks" Location

It's not just about what the drone has; it's about where it flies.

If the drone flies too far north, the sensors on the south side have a bad connection.
If it flies too low, it might crash or get blocked by mountains.
If it flies too high, the signal gets weak.

The authors created a mathematical "GPS" that tells the drone exactly where to hover. It's like a conductor finding the perfect spot on stage so every musician (sensor) can be heard clearly without shouting.

5. The Result: Saving Time and Money

The paper proves that by using this "Smart Hub" approach and finding the perfect balance between brain power and internet speed, you can:

Cut costs by 20%: You don't need to buy expensive, over-the-top equipment.
Speed up rescue: The data arrives at the command center much faster, meaning help arrives sooner.

The Analogy: The Pizza Delivery

Imagine you are ordering 100 pizzas for a party, but the delivery truck is very small.

Bad Strategy: You try to fit all 100 whole pizzas in the truck. It takes forever, and the pizza gets cold (high latency).
Old Strategy: You buy a bigger truck (more bandwidth). It's expensive and still slow.
This Paper's Strategy: You hire a chef (the EIH) who rides in the truck. The chef cuts the pizzas into slices, removes the crusts (processing), and only packs the delicious slices into the truck. Now, the small truck can carry the "essence" of 100 pizzas in one trip. The chef decides exactly how many slices to cut based on how fast the truck can drive.

In a Nutshell

This paper is a blueprint for smart, efficient rescue drones. It teaches us that in an emergency, you don't just need more power; you need the right kind of power in the right place. By letting the drone do some "thinking" before sending the data, we can save lives faster and spend less money doing it.

1. Problem Statement

The paper addresses the challenge of rapid situation awareness in post-disaster emergency scenarios where terrestrial communication infrastructures are damaged or non-existent.

Context: In disaster zones (e.g., floods, earthquakes), field sensors generate large volumes of high-resolution data that must be transmitted to an emergency command center with strict low-latency requirements.
Challenge: Traditional Non-Terrestrial Networks (NTNs) using Unmanned Aerial Vehicles (UAVs) and satellites face inherent limitations in size, weight, and power (SWaP), restricting their communication and computing capabilities. Simply uploading raw data often fails to meet latency constraints.
Gap: While Mobile Edge Computing (MEC) is a potential solution, existing research primarily focuses on resource orchestration (scheduling and allocation) assuming fixed resources. There is a lack of research on resource configuration (determining the optimal amount of bandwidth, computing power, and storage to deploy) combined with location optimization for Edge Information Hubs (EIHs).
Objective: To minimize the overall resource cost (weighted sum of bandwidth, data rate, CPU frequency, and storage) while guaranteeing that data uploading latency does not exceed a specific threshold ( $T_{req}$ ).

2. Methodology

The authors propose a Joint Resource Configuration and Location Optimization scheme for an EIH-equipped UAV. The methodology involves a two-step analytical approach:

A. System Modeling

Architecture: An EIH is deployed on a UAV, acting as a relay between ground sensors and a satellite. It contains communication modules (User-to-UAV, UAV-to-Satellite), a computing module (MEC server), and storage.
Data Flow: Sensor data is split into two parts:
1. To-be-computed: Processed at the EIH to extract key information (reducing data volume).
2. To-be-uploaded: Transmitted directly or after processing.
Concurrency: The system assumes concurrent transmission and computing to minimize latency, rather than sequential processing.
Variables: The optimization targets the total configured bandwidth ( $B_{total}$ ), satellite data rate ( $R^S_{total}$ ), CPU frequency ( $F_{total}$ ), storage ( $V_{total}$ ), and the UAV's horizontal location $(h_x, h_y)$ .

B. Analytical Derivations & Solution Strategy

The authors solve the non-convex optimization problem through a series of transformations:

Optimal Data Scheduling (Theorem 1):
- They derive a closed-form expression for the optimal data splitting ratio ( $\eta_u$ ) that minimizes latency and storage requirements for fixed resources.
- This reduces the problem complexity by eliminating the scheduling variable $\eta_u$ .
Resource Configuration Optimization (Theorems 2–5):
- Reduction: By analyzing the piecewise latency functions, they prove that the optimal solution lies in a specific region where communication and computing flows are balanced (no data accumulation in storage).
- Closed-Form Solutions: They derive closed-form expressions for the optimal user-EIH bandwidth ( $B^*_u$ ) and the optimal split between satellite data rate ( $R^S_{total}$ ) and computing frequency ( $F_{total}$ ).
- Key Insight: The decision to compute data locally or upload it raw depends on a threshold condition:
  $\frac{\rho_u}{1-\zeta_u} \leq \frac{a_2}{a_3}$
  Where $\rho_u$ is computing intensity, $\zeta_u$ is the compression ratio, and $a_2/a_3$ represents the cost ratio of satellite transmission vs. computing. If computing is cost-effective, data is processed; otherwise, it is uploaded raw.
Location Optimization (Theorems 6 & Algorithm 1):
- With the optimal resource configurations expressed as functions of the UAV location, the problem reduces to finding the optimal $(h_x, h_y)$ .
- Approximation: Due to the complexity of the channel gain expectation, they use Random Matrix Theory to approximate the ergodic rate.
- SCA Method: The resulting non-convex problem is solved using Successive Convex Approximation (SCA). Non-convex constraints are replaced with first-order Taylor expansions, transforming the problem into a convex one solvable via iterative algorithms.

3. Key Contributions

Novel Framework: Proposed an EIH-empowered NTN model that jointly optimizes resource configuration (how much to deploy) and resource orchestration (how to use it), a gap in existing literature.
Closed-Form Solutions: Derived analytical, closed-form optimal solutions for communication and computing resource configuration. This reveals the fundamental trade-off between communication and computing resources and eliminates the need for heavy numerical optimization for resource sizing.
Location Algorithm: Developed an efficient iterative algorithm based on SCA to optimize the EIH's physical location, considering both sensor distribution and data volume.
Cost Efficiency: Demonstrated that optimizing the amount of resources prevents over-provisioning, significantly reducing operational costs compared to fixed-resource benchmarks.

4. Simulation Results

The authors validated their scheme against two benchmarks:

Scheme 1: Joint optimization assuming sequential (not concurrent) processing.
Scheme 2: Equal resource allocation heuristic.

Key Findings:

Cost Reduction: The proposed scheme achieves approximately 20% reduction in overall resource costs compared to Scheme 1. This is attributed to the concurrent processing assumption and optimal resource sizing.
Latency Satisfaction: The derived closed-form bandwidth ( $B^*_{total}$ ) is proven to be a necessary and sufficient condition to meet latency constraints.
Location Optimality: In various topologies, the proposed location optimization scheme outperforms a geometric center heuristic and a previous state-of-the-art scheme [32], closely matching the results of a computationally expensive brute-force search.
Trade-off Verification: Simulations confirmed that increasing bandwidth beyond the optimal $B^*$ does not yield further savings in satellite resources, validating the theoretical optimality.

5. Significance

Mission-Critical Applications: The work provides a systematic framework for deploying 6G aerial networks in disaster relief, ensuring that critical data is transmitted within strict time limits without wasting scarce resources.
6G Readiness: It addresses the "Edge Intelligence" paradigm in 6G, moving beyond simple connectivity to synergistic communication-computing design.
Practical Deployment: The derivation of closed-form solutions is highly significant for real-world implementation, as it allows for rapid, low-complexity decision-making in dynamic emergency scenarios where computational power on the UAV is limited.
Resource Synergization: It establishes a clear mathematical relationship between computing costs and communication costs, guiding network planners on when to process data at the edge versus transmitting it raw.