🔢 mathematics

HAPS as a Hypercell: Enabling Coverage and Capacity Carrier Shutdown in Cellular Networks

This paper proposes a novel "HAPS-Hypercell" architecture where High-Altitude Platform Stations act as a wide-area non-terrestrial layer to enable the shutdown of both coverage and capacity terrestrial macro-cells, thereby significantly reducing overall network energy consumption while adhering to 3GPP standards.

Original authors: Matteo Bernabè, David López-Pérez, Nicola Piovesan

Published 2026-07-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Matteo Bernabè, David López-Pérez, Nicola Piovesan

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 bustling city where the lights in every single building are kept on 24 hours a day, 7 days a week, just in case someone needs to walk through. Even at 3:00 AM when the streets are empty, the power company keeps the lights blazing. This is exactly how our current mobile networks work: they keep massive "macro-cells" (the big towers that provide coverage) running at full power all the time, even when very few people are using their phones. This wastes a huge amount of electricity.

This paper proposes a clever new way to save energy by introducing a "sky-based backup" system called a HAPS-Hypercell.

Here is the breakdown of how it works, using simple analogies:

1. The Problem: The "Always-On" Bulb

Currently, mobile networks have two types of towers:

Coverage Towers (The Big Bulbs): These are the giant macro-cells that stay on 24/7 to make sure you have a signal everywhere, even if no one is using it. They are expensive to run.
Capacity Towers (The Desk Lamps): These are smaller towers turned on only when traffic gets heavy (like during the day). They can be turned off at night to save power.

The problem is that the "Big Bulbs" (Coverage Towers) never turn off, even when the city is asleep. The paper argues that we are wasting energy because we can't shut these big towers down without leaving people without service.

2. The Solution: The "Sky Umbrella" (HAPS-Hypercell)

The authors introduce a High-Altitude Platform Station (HAPS). Think of this as a giant, solar-powered drone or balloon floating high in the sky, acting as a massive "Sky Umbrella."

The New Role: Instead of just being a backup for rural areas, this Sky Umbrella is designed to cover the entire city at once.
The Magic Trick: Because the Sky Umbrella is so wide and powerful, it can take over the job of the "Big Bulbs" on the ground.
The Result: When traffic is low (like at night), the ground-based "Big Bulbs" can be completely switched off. The Sky Umbrella covers the whole area, ensuring everyone still has a signal, but the network uses a fraction of the power.

3. How the Towers Talk to Each Other (Pairing)

The paper tests two different ways to organize this system, like two different ways to manage a team of workers:

Option A: The Direct Line (Non-Hierarchical)
Every single ground tower talks directly to the Sky Umbrella. If a ground tower wants to go to sleep, it asks the Sky Umbrella, "Can you cover my neighborhood?" If the Sky Umbrella says "Yes," the tower turns off.
- Pros: Very flexible; can turn off the most towers.
- Cons: The Sky Umbrella gets overwhelmed if too many towers ask to sleep at once, and it might wake up the wrong tower when traffic spikes.
Option B: The Chain of Command (Hierarchical)
The ground towers are organized in layers. Small towers talk to medium towers, and medium towers talk to the Sky Umbrella.
- Pros: It's more stable. The medium towers act as managers, waking up specific small towers only when absolutely necessary.
- Cons: It's slightly less flexible than Option A, so you might not save quite as much power.

4. The "Sleep and Wake" Algorithm

The paper describes a smart algorithm that acts like a building manager:

Check the Load: Every few minutes, the towers check how many people are using them.
The Sleep Decision: If a tower is empty and its "partner" (either the Sky Umbrella or a neighboring tower) can handle the few users who are there, the tower goes to sleep (shuts down).
The Wake-Up Decision: If the Sky Umbrella or a partner tower gets too crowded, it sends a signal to wake up the sleeping towers to help out.

5. What the Results Show

The researchers simulated this in a dense city environment and found:

Energy Savings: By using the Sky Umbrella to cover the city at night, they could turn off both the small "desk lamp" towers and the big "bulb" towers. This saved up to 12.5% of the network's power during low-traffic hours.
The Trade-off: The more aggressively they tried to save power (turning off more towers), the slightly lower the internet speed became for users. However, they found a "sweet spot" (a balanced approach) where they saved a lot of energy without making the internet noticeably slower.
Architecture Matters: The "Direct Line" approach saved the most power, but the "Chain of Command" approach was better at keeping internet speeds stable when the network got busy.

Summary

In short, this paper suggests that instead of keeping every ground tower running 24/7, we can use a high-altitude "Sky Umbrella" to cover the city at night. This allows us to switch off the heavy, energy-hungry ground towers, saving significant electricity while still keeping our phones connected. It's like turning off the stadium lights and using a single, powerful spotlight to illuminate the whole field when the game is over.

Technical Summary: HAPS as a Hypercell for Energy-Efficient Carrier Shutdown

Problem Statement
The transition to 5G-Advanced and 6G systems has intensified the energy consumption challenge for mobile networks. Radio Access Network (RAN) energy costs constitute 20–40% of operator operational expenditures, with networks consuming approximately 1–1.3% of global electricity annually. A primary source of inefficiency lies in the rigid network design that distinguishes between "coverage cells" (high-power macro-cells that must remain active to ensure ubiquitous service) and "capacity cells" (dynamically activated to handle traffic variations). While Carrier Shutdown (CS) mechanisms allow capacity cells to deactivate during low-load periods, coverage cells must remain active regardless of traffic load, limiting potential energy savings. Furthermore, existing Non-Terrestrial Network (NTN) efforts, including High-Altitude Platform Stations (HAPS), primarily focus on coverage extension for rural areas rather than optimizing terrestrial network energy efficiency.

Methodology
The paper proposes a novel framework where an HAPS acts as a "Hypercell," a wide-area non-terrestrial layer capable of assuming the coverage role of multiple terrestrial macro-cells. This enables the coordinated shutdown of both capacity and coverage terrestrial layers.

System Model: The authors developed a comprehensive, 3GPP-compliant system-level simulator (using the Giulia simulator) modeling a dense urban multi-layer network. The scenario includes co-located 4G LTE (2 GHz) and 5G NR (3.5 GHz) macro-cells, alongside a single-sector HAPS (1.8 GHz) located at the network center. The model incorporates realistic channel conditions (3GPP UMa statistical models, Rician fading), MIMO antenna arrays (ULA for 4G, UPA for 5G/HAPS), and dynamic traffic distributions with hotspots.
Architectures: Two HAPS-Hypercell pairing architectures were evaluated:
1. Non-Hierarchical (HAPS-NH): All terrestrial cells act as capacity nodes paired directly with the HAPS.
2. Hierarchical (HAPS-H): Terrestrial capacity cells pair with terrestrial coverage cells, which in turn pair with the HAPS. This allows intermediate coverage cells to manage local traffic and wake up deactivated capacity cells.
Algorithm: A distributed Carrier Shutdown algorithm was implemented. Cells autonomously decide to deactivate or reactivate based on local and paired-cell Physical Resource Block (PRB) loads. Shutdown occurs if the aggregate load of a cell and its pair is below a threshold ( $\theta_{shtdn}$ ); reactivation occurs when a coverage node's load exceeds a wake-up threshold ( $\theta_{wkup}$ ).
Evaluation: Simulations were conducted over 24-hour periods under varying traffic demands (Low to High) and three CS policies (Conservative, Balanced, Aggressive) defined by different threshold values.

Key Contributions

Formalization of the HAPS-Hypercell: The paper defines the HAPS-Hypercell as an energy-efficient framework that decouples coverage from terrestrial infrastructure, enabling the deactivation of both capacity and coverage terrestrial cells.
Structured Pairing Mechanisms: The authors introduce two distinct pairing architectures (Non-Hierarchical and Hierarchical) that facilitate distributed decision-making and information exchange between terrestrial layers and the NTN layer.
3GPP-Compliant System-Level Evaluation: The study provides a detailed simulation framework capturing realistic channel conditions, traffic dynamics, and MIMO operations to quantify energy savings and performance trade-offs, moving beyond simplified models used in prior work.

Results

Power Consumption: Under low-traffic conditions, the HAPS-Hypercell approach significantly outperforms terrestrial-only CS.
- In the early morning (4:00–7:00), the HAPS-NH architecture with an aggressive profile achieved a 12.5% reduction in average network power compared to a terrestrial network with no CS.
- Compared to a terrestrial network using conventional CS, the HAPS-NH aggressive profile yielded an additional 10.5% reduction in 24-hour average power consumption.
- Under high-traffic conditions, savings were reduced (e.g., 9.2% reduction vs. TN No CS in early morning) as more cells remained active, but the HAPS-NH architecture consistently outperformed HAPS-H.
QoS and Rate Distribution:
- Conservative profiles resulted in minimal impact on user data rates (mean reduction of ~1.3%), closely matching the "No CS" baseline.
- Aggressive profiles offered the highest energy savings but incurred higher rate reductions (up to 13.0% for HAPS-NH under high traffic).
- Architectural Trade-offs: The HAPS-NH architecture provided greater flexibility and energy savings but suffered from QoS degradation under high load due to centralized control at the HAPS, which could lead to suboptimal cell reactivation. The HAPS-H architecture mitigated saturation risks by utilizing intermediate coverage cells to wake up capacity cells, offering better QoS stability at the cost of slightly lower energy savings.

Significance and Claims
The paper claims to be the first to demonstrate that an NTN HAPS layer can effectively assume the coverage role of multiple terrestrial macro-cells, thereby enabling the shutdown of both capacity and coverage layers simultaneously. The authors assert that this approach reduces overall network power consumption while maintaining service availability, addressing a critical inefficiency in current network operations where coverage cells remain active regardless of load.

The study highlights that while straightforward HAPS integration yields significant gains, the choice of pairing architecture is crucial. The results suggest that distributed decision-making based solely on local PRB load has limitations, particularly in non-hierarchical setups under high load. Consequently, the paper positions this work as a foundation for future optimization, including the development of AI-based policies and advanced reasoning models to better capture network dynamics and maximize energy savings for future greener 6G networks.