Which propagation model should I use for coverage planning at different frequencies?

Free-space path loss applies only to clear line-of-sight links such as microwave backhaul. For macrocell coverage below 2 GHz, Okumura-Hata and its COST 231 extension (valid to 2 GHz) remain common for quick estimates. For 3GPP 5G work, the UMa, UMi, and RMa models cover 0.5 to 100 GHz with explicit LOS/NLOS branches and shadowing terms. Above 6 GHz, the close-in (CI) and alpha-beta-gamma (ABG) models calibrated from measurement campaigns are preferred because oxygen absorption near 60 GHz and rain attenuation become significant. For final design, ray-tracing or measurement-tuned models replace empirical formulas because building geometry dominates millimeter-wave coverage.

Link Engineering

Coverage Planning

Q: How do you calculate cell radius from a coverage planning link budget?

Start with the maximum allowable path loss (MAPL) from the link budget: MAPL = EIRP minus receiver sensitivity plus receiver antenna gain minus all margins (fade margin, body loss, interference margin, penetration loss). For a 5G mid-band site at 3.5 GHz with 60 dBm EIRP, a receiver sensitivity of minus 100 dBm, an 8 dB fade margin, and 20 dB building penetration, the MAPL is roughly 132 dB. Plug that MAPL into the chosen propagation model (Okumura-Hata, COST 231, or 3GPP UMa) and solve for distance. With UMa NLOS at 3.5 GHz, 132 dB path loss corresponds to a cell radius near 300 to 400 m in dense urban areas. Capacity often constrains the radius below this coverage-limited value, so the final inter-site distance is the smaller of the coverage and capacity limits.

Q: What coverage probability and fade margin should I design to?

Most cellular networks target 90 to 95 percent area coverage probability at the cell edge, which is the fraction of locations exceeding the minimum required signal. The fade margin needed to hit that target depends on the log-normal shadowing standard deviation (typically 6 to 10 dB in urban environments) and is derived from the Q-function. For 95 percent edge coverage with 8 dB shadowing, the required margin is about 1.64 times sigma, roughly 13 dB, though area-vs-edge integration over the cell often reduces the required edge margin by 2 to 4 dB. Public safety and mission-critical networks push to 97 to 99 percent, demanding larger margins, denser sites, or in-building distributed antenna systems.

/KUV-er-ij PLAN-ing/

Before a single site is built, engineers must predict where a network will deliver usable signal and where it will not. That predictive discipline combines a link budget with an empirical or ray-traced propagation model to map received power, signal quality, and capacity across a target service area. The output sets the maximum allowable path loss, which when fed into the model yields cell radius, the number of sites required, and the antenna heights, tilts, and azimuths needed to meet a stated coverage probability (commonly 90 to 95% at the cell edge). Coverage planning spans cellular macro and small cells, fixed microwave links, public-safety networks, and satellite footprints, balancing signal reach against capacity, interference, and cost.

Category: Link Engineering

Edge Coverage Target: 90 to 95%

Typical Fade Margin: 8 to 13 dB

From Link Budget to Site Count

The core of coverage planning is solving the link budget for the maximum allowable path loss (MAPL), then inverting a propagation model to convert that loss into a serviceable distance. The link budget accounts for transmit power, antenna gains, feeder and connector losses, and every margin that protects the link against real-world impairments: log-normal shadowing (the fade margin), building or vehicle penetration, body loss, and an interference margin that reflects loading from neighboring cells. A 3.5 GHz 5G mid-band macro site with 60 dBm EIRP, minus 100 dBm receiver sensitivity, 20 dB indoor penetration, and an 8 dB fade margin yields an MAPL near 132 dB, which the 3GPP Urban Macro model maps to a roughly 300 to 400 m radius in dense urban clutter.

Coverage is only half the problem. Each cell also has a finite capacity set by spectrum, modulation, and the number of sectors, so the planner computes both a coverage-limited radius and a capacity-limited radius and adopts the smaller of the two as the design inter-site distance. In dense urban deployments capacity usually dominates, packing sites closer than coverage alone would require; in rural macro deployments coverage dominates, and sites are spaced as far as the MAPL allows. The frequency band drives this tension sharply: sub-GHz bands propagate far but carry little capacity, while millimeter-wave bands carry enormous capacity over very short, line-of-sight-dependent ranges.

Modern planning tools replace hand calculations with terrain databases, clutter morphology layers, and ray-tracing engines, then calibrate model parameters against drive-test or crowd-sourced measurements. The empirical formulas below remain essential for sanity checks, first-pass site counts, and understanding why a predicted plot looks the way it does.

Governing Equations

Maximum Allowable Path Loss:
MAPL = P_EIRP − S_RX + G_RX − (M_fade + L_pen + L_body + M_int) dB

Free-Space Path Loss (LOS reference):
FSPL = 20·log₁₀(d_km) + 20·log₁₀(f_MHz) + 32.44 dB

Okumura-Hata (urban, 150 to 1500 MHz):
L = 69.55 + 26.16·log₁₀(f) − 13.82·log₁₀(h_b) − a(h_m) + (44.9 − 6.55·log₁₀(h_b))·log₁₀(d)

Edge Fade Margin (log-normal shadowing):
M_fade ≈ σ × Q⁻¹(1 − P_edge) (e.g. σ = 8 dB, P_edge = 95% → M ≈ 13 dB)

P_EIRP = effective isotropic radiated power, S_RX = receiver sensitivity, G_RX = receive antenna gain, h_b/h_m = base/mobile antenna heights (m), d = distance (km), f = frequency (MHz), σ = shadowing standard deviation.

Propagation Models for Coverage Prediction

Model	Frequency Range	Environment	Inputs	Typical Use
Free-Space (FSPL)	Any	Clear LOS	d, f	Microwave backhaul, satellite
Okumura-Hata	150 to 1500 MHz	Urban / suburban / rural	d, f, h_b, h_m	Legacy 2G/3G macro
COST 231 Hata	1500 to 2000 MHz	Urban / suburban	d, f, h_b, h_m	GSM 1800, early UMTS
3GPP UMa / UMi	0.5 to 100 GHz	Urban macro / micro	d, f, LOS/NLOS, σ	4G/5G system design
Close-In (CI) / ABG	> 6 GHz	mmWave urban	d, f, PLE, measured fit	5G mmWave, fixed wireless
Ray-tracing	Any	Site-specific 3D	Building geometry, materials	Final dense-urban design

Common Questions

Frequently Asked Questions

How do you calculate cell radius from a coverage planning link budget?

Solve the link budget for maximum allowable path loss (MAPL = EIRP − receiver sensitivity + receive antenna gain − the sum of margins), then invert the propagation model for distance. A 3.5 GHz 5G site with 60 dBm EIRP, minus 100 dBm sensitivity, 8 dB fade margin, and 20 dB penetration gives roughly 132 dB MAPL, which 3GPP UMa NLOS maps to a 300 to 400 m radius in dense urban areas. The final radius is the smaller of this coverage-limited value and the capacity-limited value.

What coverage probability and fade margin should I design to?

Cellular networks typically target 90 to 95% area coverage probability at the cell edge. The required fade margin follows from the shadowing standard deviation (6 to 10 dB urban) and the Q-function; for 95% edge coverage with 8 dB shadowing the margin is about 1.64σ ≈ 13 dB, though area-versus-edge integration over the cell often shaves 2 to 4 dB off the edge requirement. Public-safety and mission-critical systems push to 97 to 99%, needing larger margins or denser sites.

Which propagation model should I use at different frequencies?

Use FSPL only for clear line-of-sight links. Below 2 GHz, Okumura-Hata and COST 231 give fast macro estimates. For 4G/5G, the 3GPP UMa, UMi, and RMa models span 0.5 to 100 GHz with LOS/NLOS branches. Above 6 GHz, measurement-calibrated close-in (CI) and ABG models handle mmWave, where rain and oxygen absorption near 60 GHz matter. For final dense-urban design, ray-tracing tuned to drive-test data replaces empirical formulas.

Related Terms

← Coverage Map Coverage Prediction →