Modeling Non-Uniform User Distributions in Wireless Networks

This direction is merged into another direction: spatial demand heterogeneity. Its last version stays readable and citable here.

Version 1 of 118 papers18 conceptsfrozen 2026-10-04

In the mathematical modeling of wireless networks, there is a convenient fiction: the Homogeneous Poisson Point Process (PPP). For decades, researchers have treated mobile users as if they were gas molecules—distributed uniformly and independently across a landscape. This assumption makes the math tractable, but it misses the most fundamental characteristic of human behavior: we cluster.

Recent research has shifted toward a more realistic spatial framework that accounts for "hotspots," "repulsion," and non-uniform density. By moving beyond uniform models, researchers are discovering that our current network designs may be fundamentally misaligned with how people actually use them.

The Failure of the Uniformity Assumption

The standard uniform model assumes that the probability of finding a user at any given coordinate is constant. However, as S. Kishore, S.C. Schwartz, and L.J. Greenstein pointed out in Uplink throughput in a single-macrocell/single-microcell CDMA system with application to data access points, real-world users naturally gravitate toward "Data Access Points" (DAPs).

When users cluster, the interference dynamics change. In a uniform model, interference is smoothed out. In reality, a dense cluster of users creates a localized "interference storm." Tian Ding, David López‐Pérez, and colleagues demonstrated in Uplink Performance Analysis of Dense Cellular Networks with LoS and NLoS Transmissions that ignoring these hotspots leads to a significant underestimation of interference. They found that in dense 5G networks, the traditional strategy of increasing user power compensation actually backfires because it exacerbates the interference within these clusters.

Modeling Repulsion and Attraction

To fix these models, researchers are introducing two competing spatial concepts: repulsion for base stations and attraction for users.

  1. Base Station Repulsion: Unlike users, base stations are not placed randomly; engineers intentionally space them out to maximize coverage. Chang‐Sik Choi, Jae Oh Woo, and Jeffrey G. Andrews proposed a hybrid model in An Analytical Framework for Modeling a Spatially Repulsive Cellular Network that combines structured grids with random processes to reflect this "intentional spacing."
  2. User Attraction: Users gather around landmarks. In indoor environments, Zhi Chai and Justin P. Coon noted in Improved Model and Analysis for RIS-Assisted Indoor Terahertz Wireless Networks that people cluster around desks or kiosks.

When you combine these—structured base stations serving clustered users—the "coverage probability" predicted by old models falls apart. The network's actual performance is dictated by the specific geometry of these hotspots rather than the average density of the whole cell.

Why Spatial Realism Matters for 6G

This shift in modeling isn't just a theoretical exercise; it has immediate consequences for emerging technologies like UAV-aided networks and millimeter-wave (mmWave) communications.

In A Unified Spatial Framework for UAV-Aided MmWave Networks, Wenqiang Yi and Yuanwei Liu show that because mmWave signals are easily blocked, the exact spatial distribution of users determines whether a UAV should be deployed at a specific "optimal altitude." If users are clustered, the UAV can hover directly over the hotspot to maintain Line-of-Sight (LoS), a strategy that a uniform model would fail to prioritize.

Similarly, Walid Saad and Mingzhe Chen argue in Echo-Liquid State Deep Learning for 360° Content Transmission and Caching in Wireless VR Networks that predicting where users will move and cluster allows for proactive content caching. If a network knows a crowd is forming, it can pre-load VR data to local edge servers, improving reliability by over 25%.

The Localization Challenge

Finally, non-uniformity affects how devices find themselves. Maximilian Stahlke and Christopher Mutschler found that indoor localization algorithms often fail because they assume users move everywhere equally. In Velocity-Based Channel Charting with Spatial Distribution Map Matching, they demonstrate that by acknowledging that users only traverse specific paths (and avoid walls or furniture), they can align radio "fingerprints" to floor plans with zero manual labeling.

By embracing the "messiness" of human spatial patterns—the clusters, the paths, and the landmarks—researchers are building a more robust foundation for the next generation of wireless connectivity. The goal is no longer to provide uniform coverage for a uniform world, but to provide targeted, high-capacity access where people actually stand.

Papers in this version

History & citationVersion 1 · 2026-10-04

Merged into another direction · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_modeling_non_uniform_user_distributions__d4f3da46,
  title = {Modeling Non-Uniform User Distributions in Wireless Networks},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/modeling-non-uniform-user-distributions-in-wireless-networks/rdr_d4f3da46471346709e97c2b783c51ad7},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_modeling_non_uniform_user_distributions__d4f3da46_v1,
  title = {Modeling Non-Uniform User Distributions in Wireless Networks},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/modeling-non-uniform-user-distributions-in-wireless-networks/rdr_d4f3da46471346709e97c2b783c51ad7/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}