Propagation Modeling

Derived from extensive statistical analysis of field measurements rather than analytical physics. These models use closed-form equations with curve-fitted parameters for specific environments.

• Key examples: Okumura, Hata, COST-231 Hata
• Inputs: Frequency, distance, antenna heights, environment type (urban/suburban/rural)
• Advantage: Fast computation with minimal geographic data
• Limitation: Valid only within the parameter ranges and environments of the original measurement campaigns
• Typical error: 10-15 dB standard deviation

Solve Maxwell's equations or geometric approximations directly using high-resolution terrain and building data. These physics-based approaches simulate individual ray paths including reflections, diffractions, and scattering.

• Key examples: Ray tracing, Ray launching, Finite-Difference Time-Domain (FDTD)
• Inputs: 3D city models, Digital Elevation Models, material permittivity
• Advantage: Site-specific accuracy with explicit multipath component identification
• Limitation: Computationally intensive; requires detailed geospatial databases
• Use case: Small-cell urban deployment planning and REM validation

Combine analytical physical foundations with empirical correction factors derived from measurements. These hybrid models balance computational efficiency with improved accuracy over purely statistical approaches.

• Key examples: Longley-Rice (ITM), COST-231 Walfisch-Ikegami, SUI models
• Mechanism: Apply diffraction theory to terrain profiles, then calibrate with measured loss coefficients
• Advantage: Terrain-aware without full 3D building data requirements
• Limitation: Cannot resolve individual multipath components
• Primary use: Macro-cell coverage prediction for spectrum cartography

Model the wireless channel as a random process characterized by probability distributions rather than deterministic path calculations. Essential for capturing fading statistics in dynamic environments.

• Key components: Path loss exponent, shadow fading variance (log-normal), multipath fading distributions (Rayleigh, Rician, Nakagami)
• Outputs: Probability of outage, coverage reliability contours, fade margins
• Advantage: Quantifies uncertainty and provides statistical guarantees for link budgets
• Integration: Often layered on top of empirical or deterministic path loss predictions
• Critical for REM: Generates confidence intervals for spectrum opportunity maps

Leverage neural networks and geostatistical methods to learn propagation characteristics directly from crowdsourced measurements or sensor networks without explicit physics equations.

• Key techniques: Gaussian Process Regression, Graph Neural Networks, Deep Neural Networks
• Inputs: Sparse RF sensor data, satellite imagery, land-use classification maps
• Advantage: Adapts to environments where traditional models fail; provides native uncertainty quantification
• Limitation: Requires substantial training data; generalization across frequencies remains challenging
• REM application: Enables real-time REM updates from distributed spectrum sensors via Kriging interpolation and federated learning

The foundational reference model assuming unobstructed line-of-sight propagation in a vacuum. All other models build upon or deviate from this ideal Friis transmission equation.

• Formula: FSPL = 32.45 + 20log₁₀(f_MHz) + 20log₁₀(d_km) [dB]
• Characteristic: Signal power decays with the square of distance (n=2 path loss exponent)
• Role: Serves as the lower-bound reference for all propagation predictions
• Deviation: Real-world models add excess loss terms for clutter, diffraction, and atmospheric effects
• Validation check: Any model predicting less loss than free space is physically impossible

A bare-earth 3D raster representation of terrain surface topography used as a critical input layer for calculating diffraction loss and line-of-sight obstructions in RF propagation prediction tools.

• Provides elevation data at regular grid postings (e.g., 30m SRTM)
• Essential for Knife-Edge Diffraction calculations
• Combined with clutter data for complete path profile analysis
• Resolution directly impacts propagation prediction accuracy
• Serves as the foundational layer in Radio Environment Maps

A spatial layer within a REM that models the large-scale, log-normal signal variation caused by macroscopic obstructions between the transmitter and receiver, distinct from distance-dependent path loss.

• Characterized by a standard deviation (typically 6-12 dB)
• Exhibits spatial correlation over tens of meters in urban environments
• Modeled as a zero-mean Gaussian random variable in dB
• Critical for calculating coverage probability and cell edge performance
• Often generated using Kriging Interpolation from sparse measurements

A detailed digital representation of urban geometry, including building footprints, heights, and material properties, used as a geometric database for ray-tracing propagation engines to simulate urban small-cell coverage.

• Enables deterministic prediction of reflection and diffraction paths
• Material properties define complex permittivity for accurate reflection coefficients
• Level of Detail (LoD) ranges from simple extrusions to detailed facades
• Essential for mmWave and sub-THz propagation modeling
• Integrated with RF Digital Twins for real-time network optimization