Clustered UAV Networks With Millimeter Wave Communications: A Stochastic Geometry View

Wenqiang Yi, Yuanwei Liu, Yansha Deng, Arumugam NallanathanIEEE Trans. Communications 2020 · DOI

Abstract

This paper presents a rigorous mathematical framework for analyzing Unmanned Aerial Vehicle (UAV) networks utilizing millimeter-wave (mmWave) frequencies to provide high-speed data to clustered users. By leveraging stochastic geometry, the study demonstrates that optimizing UAV altitude and antenna array size enables these aerial networks to significantly outperform traditional terrestrial base stations in both coverage and spectral efficiency.

Introduction

The escalating demand for high-bandwidth applications, such as 4K video streaming and emergency response data transmission, has positioned UAVs as a promising solution for providing temporary, high-capacity wireless coverage. While mmWave frequencies offer the necessary bandwidth, they are highly susceptible to blockages and experience significant path loss. Existing research often relies on simplified 2D models or neglects the natural clustering of users. This work addresses these gaps by proposing a tractable 3D spatial model that incorporates realistic 3D antenna patterns, UAV altitude, and user "hot spots."

System Model and Methodology

The authors employ stochastic geometry to model network performance, avoiding the need for exhaustive site-specific simulations. The network is modeled as a Poisson Cluster Process (PCP), specifically a Thomas cluster process, where UAVs (parent points) hover over the centers of user clusters (daughter points).

3D system model of a clustered UAV network with building blockages and user clusters.

A critical component of this model is the 3D blockage environment, which uses a stochastic approach to determine the probability of Line-of-Sight (LOS) versus Non-Line-of-Sight (NLOS) transmissions based on urban building densities. To serve users, two selection schemes are evaluated: Random Selection (RS) and the $i$-th Nearest (IN) scheme.

Furthermore, the work implements a Uniform Planar Array (UPA) for 3D antenna modeling. This allows for a detailed analysis of the impact of both azimuth and elevation angles on beamforming gain, moving beyond simplified sector models.

Uniform planar array (UPA) deployment and its corresponding normalized antenna gain pattern.

The theoretical derivations are validated using Monte Carlo (MC) simulations, with environmental parameters for 28 GHz, 38 GHz, and 60 GHz sourced from established empirical measurements.

Key Insights and Results

The analysis reveals several critical findings for UAV network design:

  • Optimal Altitude: There exists a "sweet spot" for UAV altitude. Flying too low increases blockage from ground obstacles, while flying too high increases distance-related path loss. The model demonstrates that UAVs at an optimal altitude can outperform fixed 30m terrestrial base stations.
  • NLOS Impact: In mmWave UAV systems, NLOS signals are so heavily attenuated that they can be effectively ignored in performance calculations without sacrificing accuracy, significantly simplifying network design.
  • Interference Dynamics: As UAV density increases, the system transitions from being noise-limited to interference-limited, where signals from neighboring UAVs become the primary performance constraint.
  • Antenna Scaling: Increasing the number of antenna elements ($M$) is a powerful mechanism to compensate for the high path loss at higher frequencies like 60 GHz.

The authors successfully derived tractable closed-form expressions for Coverage Probability and Area Spectral Efficiency (ASE). Numerical results confirm that these theoretical approximations match Monte Carlo simulations with high precision. While 28 GHz is robust, 60 GHz can achieve superior performance if the antenna array is sufficiently large to provide high beamforming gain.

Conclusion and Future Work

This research provides a foundational framework for the deployment of mmWave UAV networks. While the current analysis assumes a fixed altitude for UAVs and focuses on the downlink, it sets the stage for future explorations. Suggested future directions include modeling altitude distributions, addressing uplink and backhaul challenges, and investigating hybrid precoding for MIMO-UAV systems to further enhance capacity by combining digital and analog signal processing.

Summary updated

BibTeX
@article{yi2020clustered,
  author = {Wenqiang Yi and Yuanwei Liu and Yansha Deng and Arumugam Nallanathan},
  title = {{Clustered UAV Networks With Millimeter Wave Communications: A Stochastic Geometry View}},
  journal = {IEEE Trans. Communications},
  year = {2020},
  doi = {10.1109/tcomm.2020.2980787},
}
Figure