Performance Analysis and Experimental Validation of UAV Corridor-Assisted Networks

TL;DR

This work tackles the challenge of modeling UAV corridor networks by introducing a 1D stochastic geometry framework using a $1$D $BPP$ and a finite $1$D $HPPP$ to place UAV-BSs along a predefined aerial corridor. It couples distance-based path loss with shadowing modeled by an inverse-Gamma distribution and employs a maximum average received power association policy, deriving exact coverage probability expressions and a dominant-interferer simplification. The analytical results are validated through an air-to-ground measurement campaign in Prague, demonstrating the necessity of incorporating shadowing in the association policy and revealing that UAV height variability is well captured by a Normal distribution. The findings offer practical design guidance, showing that deploying large corridors at low heights can maximize coverage and that shadowing significantly shapes user association and network performance in urban UAV corridors.

Abstract

Unmanned aerial vehicle (UAV) corridor-assisted communication networks are expected to expand significantly in the upcoming years driven by several technological, regulatory, and societal trends. In this new type of networks, accurate and realistic channel models are essential for designing reliable, efficient, and secure communication systems. In this paper, an analytical framework is presented that is based on one-dimensional (1D) finite point processes, namely the binomial point process (BPP) and the finite homogeneous Poisson point process (HPPP), to model the spatial locations of UAV-Base Stations (UAV-BSs). To this end, the shadowing conditions experienced in the UAV-BS-to-ground users links are accurately considered in a realistic maximum power-based user association policy. Subsequently, coverage probability analysis under the two spatial models is conducted, and exact-form expressions are derived. In an attempt to reduce the analytical complexity of the derived expressions, a dominant interferer-based approach is also investigated. Finally, the main outcomes of this paper are extensively validated by empirical data collected in an air-to-ground measurement campaign. To the best of the authors' knowledge, this is the first work to experimentally verify a generic spatial model by jointly considering the random spatial and shadowing characteristics of a UAV-assisted air-to-ground network.

Figures (12)

• Figure 1: Illustration of the UAV corridor-assisted network. The serving UAV-BS is chosen by considering the shadowing conditions of UAV-BSs in the user association policy.
• Figure 2: The geometric locus of the points with received power $Pr_y = S_y d_y^{-\alpha}$, versus $S_y$ and $d_y$ given a maximum received power $s_0$
• Figure 3: Coverage probability, versus threshold $\theta$ under exact and dominant interferer approaches, for two values of $N$ UAV-BSs. Markers denote analytical results.
• Figure 4: Coverage probability versus $\lambda_{UAV}$ for different values of $h$ and $q$. Markers denote analytical results.
• Figure 5: Coverage probability versus $R$ for two values of $\theta$ and different values of $h$ and $q$. Markers denote analytical results.