When deploying wireless network nodes in the sky—whether they are Unmanned Aerial Vehicles (UAVs) acting as flying base stations, High-Altitude Platforms (HAPs), or Low Earth Orbit (LEO) satellites—the most fundamental design variable is altitude. Intuitively, it seems that flying higher is always better: a higher vantage point clears ground clutter, avoids obstacles, and expands the node's footprint.
However, the physics of air-to-ground (A2G) channels reveals a non-monotonic relationship between altitude and network performance. Elevating an aerial node triggers a critical trade-off between two opposing forces: the geometric expansion of Line-of-Sight (LoS) channels and the simultaneous degradation of signal quality due to path loss and aggregate network interference.
The Geometry of the Air-to-Ground Channel
To understand why an optimal altitude exists, we must look at how altitude alters signal propagation. As detailed by Aziz Altaf Khuwaja et al. in their comprehensive survey, A Survey of Channel Modeling for UAV Communications, ascending allows the direct signal path to clear the first Fresnel zone. Once the link is clear of man-made and natural ground clutter, the channel behaves like an ideal free-space environment with a low path loss exponent. Furthermore, Qiuming Zhu et al. demonstrate in Geometry-Based Stochastic Line-of-Sight Probability Model for A2G Channels Under Urban Scenarios that higher altitudes fundamentally extend the "breakpoint distance," drastically increasing the probability of establishing a stable LoS connection over larger horizontal distances.
Yet, this geometric clarity comes at a steep physical cost. Received signal power is inversely proportional to the square of the 3D straight-line distance ($d = \sqrt{h^2 + r^2}$, where $h$ is altitude and $r$ is horizontal distance).
Higher Altitude (h_high) Lower Altitude (h_low)
[ Aerial Node ] [ Aerial Node ]
/ \ / \
/ \ / \
/ \ / \
/ \ / Blocked \
/ \ / by \
/ \ / Obstacle \
[Ground User 1] [Ground User 2] [Ground User 1] [Ground User 2]
(Clear LoS, but high path loss) (Stronger signal, but high NLoS risk)
As a node climbs, the physical distance increases, causing free-space path loss to dominate. In Efficient Deployment of Multiple Unmanned Aerial Vehicles for Optimal Wireless Coverage, Mohammad Mozaffari et al. formalize this mathematical sweet spot: if the node flies too low, the signal is blocked by terrestrial obstacles (Non-Line-of-Sight, or NLoS); if it flies too high, the distance attenuates the signal entirely. This creates a distinct, unimodal performance curve where coverage probability peaks at a specific altitude determined by the antenna's beamwidth and the target ground area. This non-monotonic behavior is a universal constraint, appearing across diverse architectures including millimeter-wave (mmWave) networks (Wenqiang Yi et al. in A Unified Spatial Framework for UAV-Aided MmWave Networks) and tethered aerial systems (Mustafa A. Kishk et al. in On the 3-D Placement of Airborne Base Stations Using Tethered UAVs).
The Interference Footprint Dilemma
The trade-off becomes significantly more complex when moving from an isolated link to a multi-user or spectrum-sharing network. When an aerial base station increases its altitude, its radio horizon expands. This expansion creates a severe multi-user interference problem.
In spectrum-sharing scenarios, such as an aerial base station operating over an underlaid Device-to-Device (D2D) network, a high altitude expands the drone's "interference footprint." The elevated node begins to drown out localized ground communications. To maximize overall system throughput, the aerial node must actually lower its altitude. As shown by Mohammad Mozaffari et al. in Unmanned Aerial Vehicle with Underlaid Device-to-Device Communications: Performance and Tradeoffs and Huy T. Nguyen et al. in Joint D2D Assignment, Bandwidth and Power Allocation in Cognitive UAV-Enabled Networks, reducing altitude confines the aerial signal to its primary target users, leveraging ground clutter to shield the underlaid D2D pairs from cross-tier interference.
This dynamic works in reverse for the uplink. Fernando Rosas, in Uplink performance analysis of a drone cell in a random field of ground interferers, demonstrates that as a drone base station ascends, it becomes highly vulnerable to the aggregate uplink noise generated by millions of uncoordinated ground devices. To maintain an acceptable Signal-to-Interference-plus-Noise Ratio (SINR), the system must balance its altitude against the density of the surrounding ground network—a mechanism validated via deep reinforcement learning by Ursula Challita et al. in Interference Management for Cellular-Connected UAVs: A Deep Reinforcement Learning Approach.
Downlink Complications for Aerial Users
When the drone is not the base station but the user connecting to a terrestrial cellular network, altitude introduces an entirely different set of complications. Terrestrial base stations are optimized for ground users; their antennas are physically tilted downward.
As an aerial user climbs, it flies out of the main vertical beam of the ground antennas and enters their low-gain sidelobes. Xijun Wang analyzes this in CoMP Transmission in Downlink NOMA-Based Cellular-Connected UAV Networks, showing that vertical mobility forces the receiver into highly unstable antenna regions, even though the physical path remains clear LoS.
Furthermore, high altitudes completely decouple physical proximity from signal strength. In Mobility for Cellular-Connected UAVs: challenges for the network provider, Erika Fonseca et al. reveal that at high altitudes, the geographically closest base station provides the strongest signal in only about 30% of cases. Because the line of sight is unobstructed across the horizon, the drone is simultaneously illuminated by dozens of distant, competing cell towers. As detailed by Minwei Shi et al. in The Meta Distribution of SINR in UAV-Assisted Cellular Networks and Lin Chen et al. in Dedicating Cellular Infrastructure for Aerial Users: Advantages and Potential Impact on Ground Users, this massive influx of aggregate LoS interference from distant cells causes the drone's downlink SINR to collapse as it flies higher, completely offsetting the benefits of a clearer local path.
Manifestations Across Alternative Modalities
The altitude-interference trade-off is not confined to standard radio frequencies; it manifests across different spectrums and network scales:
- Optical and Visible Light Communications (FSO/VLC): In aerial optical networks, altitude dictates the geometric angle of incidence. If a drone flies too low, the angle to a ground sensor becomes too sharp, falling outside the receiver's effective field of view. If it flies too high, atmospheric attenuation and path length degrade the laser link. The authors of Sum-Rate Maximization for UAV-assisted Visible Light Communications using NOMA: Swarm Intelligence meets Machine Learning and Yalçin Ata et al. in RIS-Embedded UAVs Communications for Multi-Hop Fully-FSO Backhaul Links in 6G Networks show that optical placement requires balancing this strict reception geometry against distance-based atmospheric fading.
- Satellite Constellations: At the orbital scale, massive LEO satellite constellations face a parallel tension. Higher orbits expand a satellite's footprint and reduce the number of routing hops needed to bypass the Earth's curvature. However, Ruibo Wang et al. demonstrate in Stochastic Geometry-Based Low Latency Routing in Massive LEO Satellite Networks that the increased propagation delay of higher altitudes outweighs the routing overhead of extra hops, making lower orbits superior for low-latency applications. Yet, lowering the orbit introduces a non-wireless penalty: Qi Wang et al. note in Integration of Navigation and Remote Sensing in LEO Satellite Constellations that lower altitudes expose satellites to heightened atmospheric drag, directly sacrificing the operational lifespan of the constellation for signal performance.
Key Foundational Work
Unmanned Aerial Vehicle With Underlaid Device-to-Device Communications: Performance and Tradeoffs (2015) — contributed 5 concept elements to this direction.
Drone Small Cells in the Clouds: Design, Deployment and Performance Analysis (2015) — contributed 4 concept elements to this direction.
Drone Small Cells in the Clouds: Design, Deployment and Performance Analysis (2015) — contributed 5 concept elements to this direction.