Knowledge Base

Challenges in Planning 5G mmWave Networks | LuxCarta

Written by LuxCarta | Apr 15, 2026 8:16:40 AM

Planning 5G mmWave networks (24 to 40 GHz bands) is fundamentally harder than planning any previous generation because the physics of millimeter-wave propagation create extreme sensitivity to physical obstructions, very short coverage distances, and catastrophic signal loss around corners or through foliage. Accurate 3D geodata, deterministic propagation modeling, and dense small cell grids are mandatory, not optional, for reliable mmWave coverage planning.

Why Is mmWave Propagation Fundamentally Different?

At millimeter-wave frequencies, electromagnetic behavior changes in ways that make all previous RF planning intuitions unreliable:

  • Free-space path loss increases dramatically: at 28 GHz, free-space path loss over 100 m is approximately 20 dB higher than at 1 GHz, severely compressing the link budget
  • Diffraction is essentially absent: at these short wavelengths, signals do not bend around building edges the way sub-6 GHz signals do; if the path is blocked, the link fails
  • Reflection and scattering dominate NLoS propagation: some coverage extends around corners via reflections, but these paths are weak, geometry-dependent, and difficult to model without accurate building data
  • Material penetration is negligible: concrete, glass, and brick absorb or reflect virtually all mmWave energy; outdoor mmWave cells do not provide meaningful indoor coverage

Challenge 1: Extreme Sensitivity to Line-of-Sight

The single largest planning challenge in mmWave networks is LoS dependency. A mmWave signal that has an unobstructed line-of-sight path can deliver multi-Gbps throughput over distances up to 150 to 300 m. The same signal with a single building or tree blocking the path delivers essentially nothing.

This creates a binary coverage landscape of LoS zones and dead zones, rather than the gradual coverage falloff that makes 4G planning forgiving. Planning tools must identify every LoS corridor with precision, which requires building footprint and height data accurate enough to correctly determine which sides of every building are within coverage and which are in shadow.

The numbers are stark: a single tree in the direct signal path at 26 GHz can impose 35.3 dB of propagation loss, equivalent to moving nearly 15 times farther away in free space. At 3.6 GHz, vegetation attenuation at the network level creates a 12.8% delta in coverage predictions across a city area.

Challenge 2: Required Small Cell Density

Because mmWave coverage distance is short (50 to 300 m depending on environment and LoS conditions), providing continuous coverage over a useful geographic area requires a density of small cells that has no precedent in cellular history. Dense urban mmWave networks in cities like New York and Tokyo have deployed cells at inter-site distances of 100 to 200 m.

Planning implications:

  • Each candidate site must be analyzed individually, with full 3D propagation modeling, not just bulk parameter sweeps
  • Site acquisition at this density is a massive operational undertaking, requiring analysis of thousands of candidate locations (street poles, building facades, traffic signal masts)
  • Geodata must cover the full area at consistent quality: a single data gap over a city block can invalidate the entire coverage plan for that block

Challenge 3: Geodata Requirements Are Unforgiving

For macro cell planning at 700 MHz, a ±5 m building height error produces a negligible change in predicted coverage. For mmWave small cell planning at 26 GHz:

  • ±2 m height error can change whether a first-floor FWA receiver is in LoS or NLoS relative to a rooftop small cell
  • Missing buildings create predicted coverage where there is actually a shadow; subscriber premises are qualified for FWA service that cannot actually be delivered
  • Missing vegetation causes systematic over-prediction of coverage along tree-lined streets, which are precisely the residential areas targeted by FWA operators

The required data specification for mmWave planning:

Layer Specification
DTM vertical accuracy 1 m RMSE ideal, 1-3 m acceptable
Building height accuracy 1 to 3 m RMSE
Building capture rate 93%+ (every missing building is a planning error)
LULC resolution 50 cm to 1 m
3D Vegetation Per-tree with trunk/canopy separation
Wall/fence data Beneficial for dense urban LoS analysis

Challenge 4: Propagation Model Calibration

Empirical propagation models calibrated for macro cell deployments are not applicable to mmWave small cells. Ray-tracing models, which explicitly compute reflection and diffraction paths from the 3D building geometry, are the most accurate approach, but they require detailed 3D building models (LOD2) and are computationally intensive.

Calibration challenges specific to mmWave:

  • Measurement campaign complexity: mmWave drive testing requires specialized equipment and produces data that covers short distances around each test point
  • Calibration transfer: factors calibrated in one street canyon do not transfer reliably to another with different building materials or geometry
  • Vegetation seasonality: foliage state (leaf-on vs leaf-off) significantly affects mmWave attenuation; models must account for this or accept seasonal prediction error

Challenge 5: Interference and Small Cell Coordination

At mmWave, the small cells are physically close to each other and to receivers, which creates complex near-field interference patterns. Planning must address:

  • Downlink inter-cell interference between adjacent small cells, particularly at sector boundaries
  • Uplink interference from nearby UEs at high transmit power in NLoS conditions
  • Backhaul planning: mmWave small cells often use wireless backhaul in the same or adjacent frequency bands, requiring additional LoS analysis for the backhaul links

Challenge 6: Mobility Handover at Small Coverage Footprints

Each mmWave cell covers a small, irregular geographic footprint. A pedestrian walking a typical city block may cross 3 to 5 cell boundaries. Handover planning requires:

  • Accurate prediction of serving cell boundaries
  • Correctly positioned overlap zones where neighboring cells provide co-coverage for handover
  • 3D building data accurate enough to identify where coverage transitions occur at street level, not just at building rooftop level

How LuxCarta Addresses This

LuxCarta provides the 3D data stack that mmWave planning requires: LOD2 building models from its deep learning extraction pipeline (90%+ accuracy, 4x productivity vs manual, per SPIE 2023 publication), per-tree vegetation data with trunk/canopy separation, and LULC at up to 50 cm resolution for precise morphology classification.

Telefónica Deutschland validated LuxCarta's data for 26 GHz 5G modeling, confirming it delivers accuracy "needed for accurate modeling of 5G networks... for Fixed Wireless Access trials." Samsung Networks Europe specifically cited the need for "sophisticated 3D maps containing geometrical and morphological information" for 5G RF planning, exactly what LuxCarta's product set delivers.

LuxCarta's IGARSS 2024 research on wall and fence extraction (achieving 80.31% precision and 86.32% recall) extends this capability to the street-level obstacle geometry that is most relevant for dense urban mmWave LoS analysis, including parapets, boundary walls, and street-level structures that fall below the building-level resolution of standard 3D city models.

The BrightEarth platform enables on-demand extraction for the specific urban zones where mmWave deployment is planned, delivering data in formats compatible with Forsk Atoll and other mmWave planning tools without a long procurement cycle.

Frequently Asked Questions

What is the realistic coverage range of a 5G mmWave small cell?

Under ideal LoS conditions in an open urban environment, a 28 GHz small cell with a high-gain antenna and standard transmit power can cover 200 to 400 m. In practical dense urban deployments with buildings and tree shadows, the usable coverage footprint is often 50 to 150 m radius, and the coverage boundary is irregular, following LoS corridors rather than forming a circle.

Does rain or weather affect mmWave 5G coverage?

Rain attenuation at 28 GHz is approximately 0.05 to 0.1 dB/m at moderate rain rates (25 mm/hr) in a typical temperate climate. For the short link distances used in mmWave small cells (50 to 200 m), rain adds 5 to 20 dB of path loss under heavy rain conditions, a meaningful effect that must be included in link budget planning as a fade margin. At 26 GHz, rain attenuation is slightly lower than at 28 GHz.

Can 5G mmWave signals reach indoor users?

Outdoor 5G mmWave cells provide negligible indoor penetration. Concrete walls impose 20 to 40 dB of loss at mmWave frequencies; glass varies widely from 3 dB (thin single-pane) to 30+ dB (energy-efficient low-E glass). Indoor mmWave coverage requires either dedicated indoor systems (DAS, small cells) or receiving antennas mounted at windows or building exteriors, relaying signals to indoor CPE.

What propagation model is most accurate for mmWave planning?

Deterministic ray-tracing models that explicitly compute reflection, diffraction, and scattering paths from 3D building geometry are the most accurate approach for mmWave planning. For city-scale planning where ray-tracing is computationally expensive, hybrid approaches use ray-tracing for detailed hotspot analysis and calibrated empirical models (3GPP UMi or TR 38.901) for broad area assessments, validated with ray-tracing results.

How many small cells per km² are needed for continuous mmWave coverage?

Continuous outdoor mmWave coverage in a dense urban environment typically requires 30 to 100+ small cells per km², depending on building density, required coverage reliability, and target speed tier. This compares to 1 to 5 macro cells per km² for 4G LTE coverage. The density requirement is why mmWave 5G is economically viable primarily in very high-footfall commercial districts where revenue per km² justifies the deployment density.

LuxCarta provides AI-powered 3D geospatial data solutions for telecom, simulation, and smart city applications worldwide. Learn more at luxcarta.com or explore on-demand extraction at BrightEarth.