Volumetric Radio Propagation and 3D Network Modeling

This direction is split into other directions: elevation-blind channel modeling. Its last version stays readable and citable here.

Version 1 of 195 papers104 conceptsfrozen 2026-10-04

For decades, wireless network planning and spatial analysis have relied on a "flat world" assumption. Whether modeling signal decay in a city or placing sensors in a factory, researchers typically treated the environment as a two-dimensional plane. However, the rise of Unmanned Aerial Vehicles (UAVs), millimeter-wave (mmWave) 5G/6G, and complex multi-story urbanism has rendered these 2D abstractions obsolete.

A growing body of research is shifting toward volumetric modeling, treating the third dimension not as a minor correction factor, but as a fundamental axis of signal dynamics and spatial logic.

The Failure of the 2D Plane

Traditional 2D models assume that users and base stations exist at a uniform height, or that vertical variations are negligible. As Luoquan Hu and Yifan Chen argue in Path Loss Models Based on Stochastic Rays, these abstractions fail to capture the complexities of multi-story buildings and varied terrain. In a 2D model, a signal might appear to have a clear path to a receiver, but in reality, it may be blocked by a floor slab or a specific building tier that exists only at a certain altitude.

This "dimensionality gap" is particularly acute in aerial networks. Tianwei Hou and Yuanwei Liu demonstrate in Multiple Antenna Aided NOMA in UAV Networks that the spatial distribution of interference in 3D is fundamentally different from 2D. In the air, a drone’s "sphere of influence" is volumetric; its signal interacts with ground users, other aircraft, and high-rise infrastructure simultaneously, creating interference patterns that 2D planar models cannot predict.

Vertical Variability and Shadowing

The core intuition behind 3D modeling is that altitude is a primary variable of performance. Research by Francesco Beritelli and colleagues in An Advanced QOS Analysis and Evaluation Method shows that mobile internet Quality of Service (QoS) varies significantly across different altitudes. A drone flying at 50 meters may experience a completely different interference profile than one at 100 meters, even if their horizontal coordinates are identical.

This verticality introduces a phenomenon known as 3D shadowing. As Shuhang Zhang and Hongliang Zhang explain in their work on SpectrumNet, physical structures cast "radio shadows" across different vertical layers. A building might block a signal for a ground user but provide a clear line-of-sight for a high-altitude platform, or vice versa. Capturing these interactions requires moving from 2D pixels to 3D voxels or continuous volumetric representations.

Neural Surrogates for 3D Geometry

While 3D ray-tracing simulations are accurate, they are computationally expensive. To solve this, researchers are developing neural architectures that "understand" 3D space. Ron Levie’s RadioUNet pioneered the use of convolutional networks to estimate radio maps, but newer models like Wi-GATr by Thomas Hehn go further. By using Geometric Algebra Transformers, these models bake E(3)-equivariance—the mathematical rules of 3D rotation and translation—directly into the AI. This allows the system to predict how signals bounce off 3D surfaces and mesh angles with 20x the speed of traditional simulations.

Similarly, Xiucheng Wang and Qiming Zhang introduced RadioDiff-3D, a generative diffusion model that treats radio propagation as a volumetric generation task. Instead of predicting a single pathloss value, it generates a full 3D field of signal characteristics, accounting for the complex multipath environments of 6G.

Applications in Complex Environments

The necessity of 3D modeling extends beyond urban cell towers:

  • Industrial and Maritime: Bin Cao demonstrates in 3-D Multiobjective Deployment that sensor networks in cramped engine rooms require 3D uncertain sensing models to account for massive physical obstacles that block signals in three dimensions.
  • Underwater: Abdulkadir Celik and Nasir Saeed show in Underwater Optical Wireless Relaying that aquatic networks face "volumetric geometric losses" and alignment challenges that 2D models simply cannot represent.
  • UAV Logistics: Nguyen Duc Minh Quang and Chang Liu developed 3D-DRM, which uses Vision Transformers to predict dynamic 3D radio maps, ensuring drones maintain connectivity as they navigate through the complex vertical corridors of a smart city.

By treating the environment as a volume rather than a surface, this research direction provides the mathematical and computational tools necessary for the next generation of autonomous, high-frequency, and multi-layered communication systems.

Papers in this version

Showing 60 of 95 papers.

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_volumetric_radio_propagation_and_3d_netw_0a6f15ac,
  title = {Volumetric Radio Propagation and 3D Network Modeling},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/volumetric-radio-propagation-and-3d-network-modeling/rdr_0a6f15ac822c4c4aaeac48611c5e310c},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_volumetric_radio_propagation_and_3d_netw_0a6f15ac_v1,
  title = {Volumetric Radio Propagation and 3D Network Modeling},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/volumetric-radio-propagation-and-3d-network-modeling/rdr_0a6f15ac822c4c4aaeac48611c5e310c/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}