How does Cloud RF compare to desktop RF planning tools?

Desktop tools like Atoll (Forsk), ASSET (TEOCO), and EDX SignalPro offer more sophisticated models (3D ray tracing, Monte Carlo simulation, traffic modeling) but require expensive licenses (10,000 to 50,000 USD per seat), powerful workstations, and local terrain database storage (100+ GB). Cloud RF provides web-based access with pay-per-use pricing (10 to 200 USD per month), no software installation, and terrain data hosted in the cloud. API access enables automation that desktop tools handle through scripting. For large operator-scale planning (thousands of sites with traffic optimization), desktop tools remain superior. For rapid feasibility studies, small network design (10 to 100 sites), IoT deployment planning, and field engineering, Cloud RF offers faster time-to-result at lower cost. Many engineers use Cloud RF for initial planning and desktop tools for detailed optimization.

Propagation & Channels

Cloud RF

Q: What propagation models does Cloud RF support?

Cloud RF implements several standard propagation models. ITU-R P.1812 is a point-to-area model for terrestrial services from 30 MHz to 6 GHz, accounting for terrain diffraction, tropospheric scatter, and ducting for distances up to 10,000 km with time variability (1% to 50% of time). Longley-Rice (Irregular Terrain Model, ITM) covers 20 MHz to 20 GHz with statistical mode for area predictions and point-to-point mode using terrain profiles. ITU-R P.526 handles diffraction over obstacles using knife-edge, Bullington, and Deygout methods for line-of-sight and near-LOS paths. Okumura-Hata and COST 231 provide empirical urban/suburban models for cellular planning. Each model has different accuracy and applicability ranges, and Cloud RF allows users to select the most appropriate model for their frequency, environment, and distance.

Q: How accurate are cloud-based propagation predictions?

Prediction accuracy depends on terrain data resolution, clutter database quality, and model selection. With 30-meter SRTM terrain data and basic land-use clutter, median prediction error is typically 8 to 12 dB standard deviation compared to drive test measurements. High-resolution LiDAR terrain (1 to 2 meter) with detailed building databases reduces this to 4 to 8 dB. Ray-tracing models (not typically available in cloud platforms due to compute cost) achieve 2 to 4 dB in dense urban environments. The largest error sources are building penetration loss variability (10 to 25 dB range), foliage loss uncertainty (seasonal and species-dependent), and indoor propagation where wall materials are unknown. For network planning, predictions within 6 to 8 dB are generally sufficient since link budgets include 8 to 15 dB fade margin.

/klowd ahr-eff/

Cloud RF is a cloud-based radio propagation modeling platform that calculates coverage, path loss, interference, and link budgets using high-resolution terrain data (SRTM, LiDAR) and land-use clutter databases. It implements ITU-R P.1812, P.526, Longley-Rice (ITM), and Okumura-Hata models, computing diffraction loss over irregular terrain. API access enables automated network planning, batch site analysis, and GIS integration for cellular, IoT LPWAN, public safety, and fixed wireless applications.

Category: Propagation & Channels

Frequency: 20 MHz to 20 GHz

Accuracy: 4 to 12 dB std. dev.

Understanding Cloud RF

Radio coverage prediction is fundamental to network planning. Before deploying a base station, engineers need to know the coverage area, signal strength at building boundaries, and interference with existing sites. Traditional RF planning tools run on dedicated workstations with locally stored terrain databases, requiring expensive licenses and specialized hardware. Cloud RF shifts this workflow to web-based access, hosting terrain elevation data (SRTM at 30 m resolution globally, LiDAR at 1 to 2 m where available) and clutter databases in the cloud, eliminating local storage and compute requirements.

The platform's propagation engine evaluates signal strength at each point in a grid by computing free-space path loss, terrain diffraction (using Bullington or Deygout methods along the terrain profile between transmitter and receiver), atmospheric absorption, and clutter loss (building, forest, suburban categories each with empirical attenuation values). For a typical macro cell coverage calculation at 1,800 MHz with 100 m grid resolution over a 10 km radius, the engine evaluates approximately 31,000 grid points, each requiring a terrain profile extraction and propagation model computation. Cloud parallelization completes this in 2 to 10 seconds versus 30 to 120 seconds on a desktop, with the ability to run hundreds of sites simultaneously for network-wide interference analysis.

Propagation Model Equations

Free-Space Path Loss:
FSPL = 32.45 + 20log(f_MHz) + 20log(d_km) (dB)

Knife-Edge Diffraction Loss:
L_d = 6.9 + 20log(√((ν - 0.1)² + 1) + ν - 0.1) for ν > -0.78

Received Power:
P_rx = P_tx + G_tx - L_cable - FSPL - L_diff - L_clutter + G_rx (dBm)

Where ν = Fresnel-Kirchhoff diffraction parameter = h√(2/(λd₁d₂/(d₁+d₂))), h = obstacle clearance height, d₁,d₂ = distances to obstacle, λ = wavelength. FSPL at 1800 MHz, 1 km = 105.5 dB.

Propagation Model Comparison

Model	Frequency Range	Distance	Accuracy (std dev)	Best For
ITU-R P.1812	30 MHz to 6 GHz	Up to 10,000 km	6 to 10 dB	General terrestrial planning
Longley-Rice (ITM)	20 MHz to 20 GHz	1 to 2,000 km	8 to 12 dB	Rural, irregular terrain
ITU-R P.526	30 MHz to 30 GHz	LOS + near-LOS	4 to 8 dB	Diffraction over obstacles
Okumura-Hata	150 MHz to 1.5 GHz	1 to 20 km	8 to 12 dB	Urban/suburban cellular
COST 231 Walfisch-Ikegami	800 MHz to 2 GHz	0.02 to 5 km	6 to 10 dB	Dense urban micro cells

Common Questions

Frequently Asked Questions

What propagation models does Cloud RF support?

ITU-R P.1812 for general terrestrial planning (30 MHz to 6 GHz, up to 10,000 km). Longley-Rice (ITM) for irregular terrain (20 MHz to 20 GHz). ITU-R P.526 for knife-edge and Bullington diffraction. Okumura-Hata and COST 231 for empirical urban/suburban cellular planning. Each model suits different frequency ranges, environments, and accuracy requirements.

How accurate are cloud-based propagation predictions?

With 30 m SRTM terrain and basic clutter, median prediction error is 8 to 12 dB standard deviation versus drive tests. High-resolution LiDAR (1 to 2 m) with building databases reduces this to 4 to 8 dB. Largest error sources are building penetration variability (10 to 25 dB) and foliage loss uncertainty. For planning, 6 to 8 dB accuracy is sufficient given 8 to 15 dB fade margins.

How does Cloud RF compare to desktop planning tools?

Desktop tools (Atoll, ASSET, EDX SignalPro) offer 3D ray tracing and traffic modeling but cost $10K to $50K per seat. Cloud RF provides web access at $10 to $200/month with API automation and no terrain data storage. For large networks (1000+ sites), desktop tools are superior. For rapid feasibility, small networks, and field engineering, Cloud RF offers faster results at lower cost.

Related Terms

← Cloud RAN Telecom Cluster-Based Model →