Abstract
Despite enhancement in rural areas is one of the main requirements in next generations of wireless networks (i.e., 5G and 6G), the low expected profit prevents telecommunication providers from investing in such sparsely populated areas. Hence, cost efficient alternatives are required for extending the cellular infrastructure to these regions. A concrete mathematical model that characterizes and clearly captures the aforementioned problem might be a key-enabler for studying the efficiency of any potential solution. In this paper, a new stochastic geometry-based model is implemented in order to show the coverage spatial variation among urban, suburban, and exurban settlements. By implementing inhomogeneous Poisson point processes (PPPs), it is possible to study performance metrics where terrestrial base stations (TBSs) are clustered around the urban center while outer aerial base stations (ABSs) are uniformly distributed outside an urban exclusion zone.
Introduction: The Rural Connectivity Gap
Telecommunication providers typically deploy terrestrial base stations (TBSs) where population density is highest to maximize return on investment (ROI). This "capacity-driven" deployment creates a stark contrast: city centers are saturated with connectivity, while suburban and rural areas are left with sparse, unreliable signals. Standard mathematical tools for analyzing these networks often assume "homogeneity," treating base station distribution as uniform. This paper addresses the gap between these idealized models and the reality of spatial "unfairness," exploring how Unmanned Aerial Vehicles (UAVs) can bridge the digital divide.
System Model and Methods
The researchers employ Stochastic Geometry to build a realistic model of a "Vertical Heterogeneous Network" (VHetNet).
Key technical components include:
- Inhomogeneous Poisson Point Processes (PPPs): Unlike standard models, the density of TBSs in this study is a function of the distance from the town center, modeled as a monotonically decreasing Gaussian distribution.
- The Exclusion Zone: To optimize resources, the model introduces an "urban exclusion zone" (radius $r_e$). UAV-mounted Aerial Base Stations (ABSs) are only deployed outside this zone, where terrestrial coverage begins to fail.
- Channel Modeling: The study uses Nakagami-m fading for small-scale signal fluctuations and incorporates specific Line-of-Sight (LoS) and Non-Line-of-Sight (NLoS) probabilities for air-to-ground links.
- Association Policy: Users connect to the base station (terrestrial or aerial) providing the strongest average received power.
Key Insights: Bridging the Divide
The analysis reveals several critical insights into rural network design:
- The "Local Minimum" Problem: Without UAV assistance, there is a sharp drop-off in coverage at the boundary where the urban core ends but the rural environment has not yet become "uniformly sparse." This "second regime" is where users suffer the most.
- Efficiency of Sparse Aerial Layers: A surprising finding is that a very low density of ABSs (e.g., 0.15 ABSs/km²) can almost entirely "even out" the coverage probability across the entire region.
- The Interference Trade-off: While adding drones helps, there is a point of diminishing returns. Eventually, interference from a high density of aerial nodes outweighs the benefit of proximity. This saturation point occurs around 0.2 ABSs/km² in the studied scenario.
- The Importance of the Exclusion Zone ($r_e$): The radius $r_e$ is a critical design lever. If too small, drones interfere with efficient urban TBSs; if too large, the "local minimum" at the urban fringe remains unaddressed.
Results and Validation
The study demonstrates significant coverage equalization. In a scenario with an 8 km exclusion zone, adding just 0.15 ABSs/km² raised rural coverage probability (20-30 km from center) from near-zero to levels comparable to the urban core (~0.7 to 0.8). Performance saturates at an ABS density of approximately 0.2 units/km², reaching a plateau of about 74% coverage probability. These analytical approximations showed a near-perfect match with Monte Carlo simulations (100,000 iterations), validating the model's utility for rapid network planning.
Future Directions
The paper identifies several areas for future research, including dynamic altitude optimization for UAVs, addressing the "backhaul" bottleneck in rural areas, and exploring spectrum sharing strategies (such as TV White Space) to further reduce interference between terrestrial and aerial tiers.
Reference: Matracia, M., Kishk, M. A., & Alouini, M.-S. (2021). Coverage Analysis for UAV-Assisted Cellular Networks in Rural Areas. IEEE Open Journal of the Vehicular Technology Society, 2, 194-206.
Summary updated