Probabilistic Caching for Small-Cell Networks With Terrestrial and Aerial Users

Fei Song, Jun Li, Ming Ding, Long Shi, Feng Shu, Meixia Tao, Wen Chen, H. Vincent Poor

Abstract: The support for aerial users has become the focus of recent 3GPP standardizations of 5G, due to their high maneuverability and flexibility for on-demand deployment. In this paper, probabilistic caching is studied for ultra-dense small-cell networks with terrestrial and aerial users, where a dynamic on-off architecture is adopted under a sophisticated path loss model incorporating both line-of-sight and non-line-of-sight transmissions. Generally, this paper focuses on the successful download probability (SDP) of user equipments (UEs) from small-cell base stations (SBSs) that cache the requested files under various caching strategies. To be more specific, the SDP is first analyzed using stochastic geometry theory, by considering the distribution of such two-tier UEs and SBSs as Homogeneous Poisson Point Processes. Second, an optimized caching strategy (OCS) is proposed to maximize the average SDP. Third, the performance limits of the average SDP are developed for the popular caching strategy (PCS) and the uniform caching strategy (UCS). Finally, the impacts of the key parameters, such as the SBS density, the cache size, the exponent of Zipf distribution and the height of aerial user, are investigated on the average SDP. The analytical results indicate that the UCS outperforms the PCS if the SBSs are sufficiently dense, while the PCS is better than the UCS if the exponent of Zipf distribution is large enough. Furthermore, the proposed OCS is superior to both the UCS and PCS.

Introduction

As 5G networks evolve, the integration of Unmanned Aerial Vehicles (UAVs) or Aerial Users (AUs) alongside traditional Terrestrial Users (TUs) presents new challenges for network management. The explosion of mobile video traffic necessitates efficient edge caching at Small-Cell Base Stations (SBSs). However, the distinct signal propagation characteristics of drones—which often have a clear Line-of-Sight (LoS) but face different interference patterns—require a more nuanced approach than traditional ground-only models.

System Model and Methods

To address these challenges, the researchers developed a mathematical framework using Stochastic Geometry. They modeled the locations of SBSs, TUs, and UAVs as independent Homogeneous Poisson Point Processes (HPPPs). A critical feature of this model is the dynamic on-off architecture, where an SBS is only active when serving a user, which accurately reflects interference in ultra-dense deployments.

System model of the ultra-dense small-cell network featuring terrestrial users (TU) and aerial users (UAV) served by small-cell base stations (SBS).

The study incorporates a multi-piece path loss model that distinguishes between LoS and Non-Line-of-Sight (NLoS) transmissions, specifically tailored to the heights of both user types (e.g., 1.5m for TUs and 30m for UAVs). The core contribution is the Optimized Caching Strategy (OCS), derived via a Lagrangian-based iterative algorithm to maximize the Successful Download Probability (SDP).

Key Results and Insights

The research highlights several critical findings regarding caching strategies:

  • The "Density Flip": In sparse networks, the Popular Caching Strategy (PCS)—caching only the most requested files—is effective. However, as SBS density increases, the Uniform Caching Strategy (UCS)—caching files randomly to increase variety—eventually outperforms PCS. In dense environments, network-wide content variety becomes more valuable than local redundancy.
  • OCS Dominance: The proposed OCS consistently outperforms both UCS and PCS by dynamically balancing popularity and variety based on network conditions.
  • UAV Performance: Aerial users generally experience a lower SDP than terrestrial users. While UAVs benefit from more LoS opportunities, the increased distance and resulting interference from a wider area of active SBSs at higher altitudes lead to signal degradation.
Average successful download probability (SDP) for (a) terrestrial users and (b) aerial users as a function of SBS density, comparing OCS, PCS, and UCS strategies.

As shown in the figures above, the OCS (solid black line) provides the upper bound for performance across all SBS densities. For TUs, the transition where UCS overtakes PCS occurs at approximately $3 \times 10^3$ SBSs/km².

Conclusion

This work provides a robust framework for optimizing content delivery in 5G networks serving both ground and aerial users. While the OCS offers clear performance gains, future research is needed to address the complexities of UAV uplink transmissions and full 3D spatial modeling to further refine these caching strategies.

Summary updated

Abstract

The support for aerial users has become the focus of recent 3GPP standardizations of 5G, due to their high maneuverability and flexibility for on-demand deployment. In this paper, probabilistic caching is studied for ultra-dense small-cell networks with terrestrial and aerial users, where a dynamic on-off architecture is adopted under a sophisticated path loss model incorporating both line-of-sight and non-line-of-sight transmissions. Generally, this paper focuses on the successful download probability (SDP) of user equipments (UEs) from small-cell base stations (SBSs) that cache the requested files under various caching strategies. To be more specific, the SDP is first analyzed using stochastic geometry theory, by considering the distribution of such two-tier UEs and SBSs as Homogeneous Poisson Point Processes. Second, an optimized caching strategy (OCS) is proposed to maximize the average SDP. Third, the performance limits of the average SDP are developed for the popular caching strategy (PCS) and the uniform caching strategy (UCS). Finally, the impacts of the key parameters, such as the SBS density, the cache size, the exponent of Zipf distribution and the height of aerial user, are investigated on the average SDP. The analytical results indicate that the UCS outperforms the PCS if the SBSs are sufficiently dense, while the PCS is better than the UCS if the exponent of Zipf distribution is large enough. Furthermore, the proposed OCS is superior to both the UCS and PCS.
BibTeX
@article{song2019probabilistic,
  author = {Fei Song and Jun Li and Ming Ding and Long Shi and Feng Shu and Meixia Tao and Wen Chen and H. Vincent Poor},
  title = {{Probabilistic Caching for Small-Cell Networks With Terrestrial and Aerial Users}},
  journal = {IEEE Trans. Vehicular Technology},
  year = {2019},
  doi = {10.1109/tvt.2019.2929839},
}
Figure