UPSim: UxNB Propagation Simulator for 3D Map-Driven FR3 Deployments

The paper introduces UPSim, a scalable, semi-deterministic propagation simulator that leverages 3D building geometry and shadow projection to generate spatially consistent FR3 air-to-ground channel maps for UAV networks, offering a computationally efficient alternative to full ray tracing while maintaining high accuracy for mobility-aware planning.

The Gist

Imagine you are trying to fly a drone equipped with a cell tower (called a "UxNB") over a busy city like Barcelona to provide internet to people on the ground. The big challenge is that buildings block the signal. Sometimes the drone has a clear view of a person (Line-of-Sight), and sometimes a skyscraper is in the way (Non-Line-of-Sight).

In the past, figuring out exactly where the signal would be strong or weak required a lot of compute to simulate millions of invisible "laser beams" (rays) bouncing off every single building. This is called Ray Tracing. It's incredibly accurate, but it's so slow and expensive that you can't use it to plan a whole city's network or track moving users on the ground under coverage from a deployed drone over a long stretch of city.

On the other hand, old-school methods just guessed the signal strength using random math. They were fast, but they didn't care about the actual shape of the buildings, so they couldn't tell you exactly where the signal would drop out as you moved down the street.

Enter UPSim.

The authors of this paper created a new tool called UPSim (UxNB Propagation Simulator). Think of UPSim as a smart "shadow-caster" that finds the perfect middle ground between the slow laser simulation and the random guesswork.

Here is how it works, using simple analogies:

1. The Shadow Puppet Show (Instead of Laser Beams)

Instead of shooting millions of laser beams from the drone to every single person on the ground, UPSim looks at the 3D map of the city and asks: "If the sun were the drone, where would the buildings cast their shadows?"

• The Analogy: Imagine holding a flashlight (the drone) high above a city. The buildings cast long, dark shadows on the ground. If you are standing in the light, you have a clear connection. If you are standing in a shadow, the building is blocking you.
• The Magic: UPSim calculates these "shadows" mathematically using a 3D map of the buildings. This is instant and doesn't require heavy compute. It instantly creates a map showing exactly which streets are "in the light" (good signal) and which are "in the dark" (blocked signal).

2. Adding the "Weather" (Calibrating the Signal)

Knowing where the shadows are is great, but it doesn't tell you how weak the signal is inside the shadow or how much it fluctuates. To fix this, the authors "taught" UPSim using data from those slow, expensive laser simulations.

• The Analogy: Imagine you know exactly where the rain clouds (shadows) are. But you also need to know if it's a light drizzle or a heavy storm inside those clouds.
• The Magic: UPSim takes the "shadow map" and adds realistic "weather patterns" to it. It uses data from the expensive laser simulations to learn how much signal is lost when flying at different heights (low vs. high) and how the signal "fades" or "flickers" as you move. It creates a complete picture of the signal quality without needing to run the slow laser simulation every time.

3. Why This Matters: The "Route" Test

The paper shows that UPSim is incredibly useful for planning ground user routes.

• The Scenario: Imagine a drone is hovering over a city, beaming internet down to people on the ground. A user walks from Point A to Point B along a street. The question is: as the user moves, where will they lose connection — for 50 metres? for 200 metres? The drone itself stays put; what changes is the ground user's position relative to the buildings.
• The Result: Because UPSim is fast, it can simulate a ground user moving along a specific street-level path under coverage from the static drone and tell you exactly how long the "outage" (loss of signal) zones along that path are.
• The Finding: They found that flying the drone higher (e.g., 150 meters up) makes the "shadows" shorter, meaning you stay in the "light" longer. However, even at high altitudes, if you fly too close to tall buildings, you still hit "dead zones."

Summary of What They Claim

• It's Fast: It uses geometry (shadows) instead of heavy physics simulations, making it scalable for big cities.
• It's Accurate: It was tested against real laser simulation data and matches it very closely.
• It's Realistic: It uses real 3D maps of Barcelona (from a global dataset called 3D-GloBFP) rather than fake, made-up city shapes.
• It's Open: The authors made the code free for anyone to use so others can build on it.

In short, UPSim is a tool that lets engineers quickly and accurately predict where a drone's internet signal will work and where it will fail in a real city, helping them plan better routes for ground users without needing heavy compute resources.

Technical Summary

Technical Summary: UPSim – UxNB Propagation Simulator for 3D Map-Driven FR3 Deployments

Problem Statement
The deployment of Uncrewed Aerial Vehicles (UAVs) carrying on-board base stations (UxNBs) in Frequency Range 3 (FR3) mid-band spectrum faces unique propagation challenges. In dense urban environments, the link budget is heavily dictated by rapid transitions between Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) conditions. While standard stochastic models (e.g., ITU or 3GPP) rely on urban layout abstractions, they often fail to capture the spatial consistency required for mobility-aware planning. Conversely, high-fidelity Ray Tracing (RT) offers geometric accuracy and continuity but incurs prohibitive computational overhead when applied to large-scale radio mapping or massive trajectory analysis. There is a need for a scalable solution that bridges the gap between computationally expensive full RT and purely stochastic channel generation, specifically leveraging the increasing availability of global 3D building data.

Methodology
The paper introduces UPSim (UxNB Propagation Simulator), a semi-deterministic, 3D map-driven solution designed for spatially consistent Air-to-Ground (A2G) propagation modeling. The methodology operates through the following stages:

1. Geometric Foundation: UPSim utilizes the 3D-GloBFP dataset, which provides consistent building heights and footprints derived from OpenStreetMap (OSM). Unlike standard OSM data where height tags are optional and inconsistent, 3D-GloBFP offers a reliable 3D description necessary for modeling blockage.
2. Geometry-Based Shadow Projection (GBSP): Instead of launching rays for every receiver position, UPSim derives deterministic visibility regions by projecting building shadows from the UxNB viewpoint onto the ground plane. This creates a total shadow map ($S_{total}$) where the complement represents the LOS region. This approach bypasses point-to-point RT for visibility determination, significantly reducing complexity.
3. Channel Synthesis: Once the LOS/NLOS state is determined geometrically, UPSim synthesizes the channel attenuation ($\Lambda$) by combining:
   • Deterministic Path Loss: A height-dependent Path Loss Exponent (PLE) for NLOS conditions, while LOS PLE is fixed.
   • Large-Scale Fading (LSF): Spatially correlated shadowing generated from filtered Gaussian noise, with standard deviation and decorrelation distances dependent on UxNB altitude and LOS state.
   • Small-Scale Fading (SSF): Modeled as a log-logistic distribution with a shape parameter dependent on the LOS/NLOS state but independent of altitude.
4. Calibration: The model parameters (PLE, LSF statistics, decorrelation distances) are calibrated against a reference FR3 Ray Tracing dataset generated over a 1km $\times$ 1km Barcelona urban scene using the 3D-GloBFP geometry.

Key Contributions
The paper outlines four primary contributions:

• Scalable Simulator: The development of UPSim, which combines deterministic shadow projection with RT-calibrated FR3 channel generation, offering a practical middle ground for large-scale analysis.
• Compact Channel Model: A unified model for path loss and spatially correlated large-scale fading as functions of UxNB altitude, integrated with state-dependent small-scale fading.
• Validation: Rigorous validation against FR3 ray tracing benchmarks using real-world geometry from the Barcelona subset of 3D-GloBFP, demonstrating accurate reproduction of empirical channel distributions.
• Route-Level Analysis: An analysis of LOS/NLOS segment lengths and outage distance statistics, proving the simulator's utility for evaluating route-level performance metrics like outage intervals.

Results
Validation results indicate that UPSim accurately reproduces the empirical Cumulative Distribution Functions (CDFs) of channel attenuation observed in the reference Ray Tracing data, outperforming the standard 3GPP TR 38.901 UMi baseline in terms of geometric consistency.

• Spatial Consistency: The simulator successfully captures the continuous evolution of channel realizations along trajectories.
• Altitude Impact: Route-level analysis reveals that increasing UxNB deployment height (e.g., from 30m to 150m) significantly increases LOS probability and reduces NLOS segment lengths. However, LOS distances remain relatively stable across heights, while NLOS zones shrink.
• Outage Statistics: Outage analysis (defined by received power thresholds) shows that outage zones are concentrated around buildings. Even at high altitudes (150m) and moderate transmit powers, significant outage probabilities (e.g., ~32% for 23 dBm EIRP) persist, highlighting the critical impact of building blockage in FR3 bands.
• Statistical Features: The CDFs of NLOS distances exhibit a "step" behavior around 50 meters, a phenomenon consistent with Manhattan-like grid structures found in the Barcelona layout.

Significance and Claims
The paper claims that UPSim offers a highly scalable, practical alternative to full ray tracing for mobility-aware planning and radio-map construction in aerial access scenarios. By decoupling the computationally intensive visibility determination (via shadow projection) from the channel synthesis (via calibrated statistical models), UPSim achieves the accuracy of ray tracing at a fraction of the complexity.

The authors emphasize that the tool enables route-based analysis of channel evolution over complex urban layouts, exposing critical trajectory-level statistics such as outage distances that are difficult to derive from purely stochastic models. The work positions itself as a step toward realistic, map-driven propagation analysis that leverages global 3D data, moving beyond the limitations of abstract urban models. The framework is provided as an open-source tool to facilitate further research and reproducibility in the community.