Propagation & Channels

Coverage Map

Q: What signal-strength threshold defines the edge of a coverage map?

The coverage edge is set by the receiver sensitivity plus a fade margin and any required signal-to-interference margin. For an LTE handset the design threshold is typically around -110 to -118 dBm RSRP for cell-edge service, while a 5G NR fixed-wireless link may require -90 dBm or better to sustain high-order modulation. A microwave point-to-point link is usually planned to a received power above its threshold by a 20 to 40 dB fade margin to meet 99.999 percent availability. The chosen threshold appears as the outermost colored contour: cells receiving power at or above it are shaded as covered, and cells below it are left uncovered. Designers usually plot several contours (for example -85, -95, -105 dBm) so the map shows graceful service degradation rather than a single binary boundary.

Q: How accurate is a predicted coverage map compared with drive-test measurements?

A coverage map built from an empirical model such as Okumura-Hata or COST 231 typically shows a prediction standard deviation of 6 to 10 dB against drive-test data in mixed terrain, dominated by unmodeled clutter and shadowing. Calibrating the model to local measurements (tuning the slope and intercept of the path-loss curve) can reduce that to 4 to 7 dB. Ray-tracing or dominant-path models using high-resolution building data reach 3 to 6 dB in dense urban areas where diffraction over rooftops and street-canyon guiding dominate. Because the error is roughly log-normal, planners add a location-variability margin: for 90 percent cell-edge coverage probability with an 8 dB shadowing sigma, about 10 dB of extra margin is folded into the threshold before the map is drawn.

Q: What input data and resolution does a coverage map require?

The core inputs are a digital terrain model (DTM) for path profile and diffraction, a clutter or land-use layer for morphology-dependent loss, and the site parameters: antenna height, azimuth, downtilt, gain pattern, EIRP, and frequency. Terrain resolution of 30 m (SRTM) is adequate for rural macrocell maps, but urban small-cell and millimeter-wave maps need 1 to 5 m building vector or lidar data because diffraction and blockage change over a few meters. The output raster is usually computed on a 10 to 50 m grid for macro networks and 1 to 5 m for indoor or street-level prediction. Higher resolution sharply increases compute time because each bin requires a separate path-loss evaluation along its profile to the transmitter.

/KUV-er-ij map/

A geographic plot of predicted or measured received signal strength that shows where a transmitter delivers usable service. Each grid cell is colored by the power a receiver would see after the link budget and a chosen propagation model account for free-space spreading, terrain diffraction, and clutter loss. The outermost contour marks the cells where received power falls to the service threshold, typically -110 to -118 dBm for LTE cell-edge handsets or higher for fixed-wireless and microwave links. Predicted maps carry a 4 to 10 dB standard deviation against drive-test data, so a location-variability margin is folded into the threshold to hit a target coverage probability of 90 to 95 percent at the cell edge.

Category: Propagation & Channels

Edge Threshold: -110 to -118 dBm (LTE)

Prediction Sigma: 4 to 10 dB

How a Coverage Map Is Computed

A coverage map turns a set of transmitter parameters into a raster of expected received power across a service area. For every grid cell the planning tool extracts a terrain profile from a digital terrain model, applies a propagation model to estimate path loss along that profile, and subtracts the result from the transmitter EIRP plus the receiver antenna gain. The cell is then shaded according to which threshold band the predicted power falls into. Because the calculation repeats independently for thousands or millions of cells, the choice of grid resolution drives both fidelity and compute time: 30 m suits rural macrocells, while 1 to 5 m is needed for street-level or millimeter-wave maps where rooftop diffraction and building blockage change over a few meters.

The propagation model is the dominant source of error. Empirical models such as Okumura-Hata and COST 231 fit measured curves to morphology classes and run quickly, but show 6 to 10 dB prediction sigma in mixed terrain. Deterministic ray-tracing and dominant-path models using high-resolution building data tighten that to 3 to 6 dB in dense urban areas at the cost of much heavier computation. Shadowing causes the residual error to follow a roughly log-normal distribution, so planners do not draw a single hard boundary at the receiver threshold; they add a margin sized to the shadowing sigma so that a stated fraction of locations at the edge actually meet service.

Predicted maps are validated and corrected with drive tests. A measurement vehicle logs RSRP or RSSI with GPS position, the data is binned to the same grid, and the model slope and intercept are tuned until prediction error is minimized. The corrected map is then used for site acceptance, interference analysis, and regulatory filings. RF Essentials supplies the antennas, low-noise front ends, and frequency converters that set the link-budget numbers feeding these maps.

Coverage Probability and Edge Margin

Received power per cell:
P_rx(dBm) = EIRP + G_rx − L_path − L_misc

Covered if:
P_rx ≥ S_rx + M_fade + M_shadow

Edge coverage probability (log-normal shadowing):
P_edge = Q( (S_rx − P̅_rx) / σ ) → M_shadow = z_p × σ

Where S_rx = receiver sensitivity, σ = shadowing standard deviation (6 to 10 dB typical), and z_p is the standard-normal quantile. Example: for σ = 8 dB and 90% edge coverage, z_p ≈ 1.28, so M_shadow ≈ 10 dB is added before drawing the threshold contour.

Prediction Model Comparison

Model	Type	Best Frequency	Prediction Sigma	Data Needed	Typical Use
Free-space (FSPL)	Theoretical	Any (LOS)	n/a (LOS only)	Distance only	Sanity check, satellite
Okumura-Hata	Empirical	150 MHz to 1.5 GHz	6 to 10 dB	Terrain + clutter class	Macrocell rural/suburban
COST 231 Hata	Empirical	1.5 to 2 GHz	6 to 9 dB	Terrain + clutter class	PCS / urban macro
Longley-Rice (ITM)	Semi-empirical	20 MHz to 20 GHz	5 to 9 dB	Terrain profile	Broadcast, point-area
Ray tracing	Deterministic	0.8 to 100 GHz	3 to 6 dB	3D building / lidar	Urban small cell, mmWave

Common Questions

Frequently Asked Questions

What signal-strength threshold defines the edge of a coverage map?

The edge contour is set by receiver sensitivity plus a fade margin and an interference margin. LTE cell-edge service is typically planned to about -110 to -118 dBm RSRP, 5G fixed-wireless needs -90 dBm or better for high-order modulation, and a microwave point-to-point link is held 20 to 40 dB above threshold for 99.999% availability. Planners plot several contours (for example -85, -95, -105 dBm) so the map shows graceful degradation rather than a single binary boundary.

How accurate is a predicted coverage map compared with drive-test measurements?

Empirical models such as Okumura-Hata or COST 231 show 6 to 10 dB prediction sigma in mixed terrain. Tuning the model to local drive-test data cuts that to 4 to 7 dB, and ray tracing with high-resolution building data reaches 3 to 6 dB in dense urban areas. Because the error is roughly log-normal, a location-variability margin is added: with an 8 dB shadowing sigma, about 10 dB of extra margin yields 90% edge coverage probability.

What input data and resolution does a coverage map require?

Core inputs are a digital terrain model for the path profile, a clutter or land-use layer for morphology loss, and site parameters: antenna height, azimuth, downtilt, gain pattern, EIRP, and frequency. 30 m terrain suits rural macrocells, but urban small-cell and millimeter-wave maps need 1 to 5 m building or lidar data. Output rasters run on a 10 to 50 m grid for macro networks and 1 to 5 m for street-level prediction, with compute time scaling with cell count.

Related Terms

← Coverage Latitude Coverage Planning →