Glossary term

Multipath Fading

Engineering definition of multipath fading covering reflected paths, destructive interference, delay spread, coherence bandwidth, Doppler and wireless validation.

Branch: Telecommunications Engineering
Glossary type: phenomenon
Content: Glossary term
Updated: Jun 26, 2026
Revision: v1.0.0 · reviewed

Definition

phenomenon

Multipath fading is variation in received signal level or quality caused by multiple delayed propagation paths combining constructively or destructively at the receiver.

Multipath fading occurs when reflections, scattering, diffraction or alternative paths create delayed signal copies. Depending on path delay, phase, motion and bandwidth, the result can be flat fading, frequency-selective fading, delay spread, inter-symbol interference, EVM degradation, packet loss or adaptive-modulation fallback.

Multipath fading is the variation in received signal level or signal quality caused by multiple propagation paths arriving at the receiver with different delays, amplitudes and phases. Reflections from buildings, ground, water, vehicles, aircraft, structures, terrain or indoor surfaces can add to or subtract from the direct signal.

The result is not just “extra path loss.” A link can have strong average received power and still fail because the channel creates deep fades, frequency notches, delay spread, inter-symbol interference or fast variation during motion. This is why wireless validation must look beyond one received-signal-strength value.

Multipath Channel Model

A simplified channel impulse response is:

h(t)=\sum_i a_i e^{j\phi_i}\delta(t-\tau_i)

where a_i is path amplitude, phi_i is phase and tau_i is path delay. The receiver sees the sum of these delayed copies. When copies arrive in phase, they can reinforce the wanted signal. When they arrive out of phase, they can create a fade.

Fade Depth

A two-path screen can show the risk. If the direct path has normalized amplitude 1 and a reflected path has amplitude a with phase difference phi, the relative received power is:

P_{rel}=|1+a e^{j\phi}|^2

For destructive interference with a=0.7 and phi=\pi:

P_{rel}=|1-0.7|^2=0.09

The fade depth relative to the direct path is:

L_{fade}=-10\log_{10}(0.09)=10.5\ \text{dB}

If the nominal link margin was 14 dB, the faded margin becomes:

M_{fade}=14-10.5=3.5\ \text{dB}

That may be below the required reserve even though the clear-condition link looked healthy.

Delay Spread

Multipath also has a timing dimension. RMS delay spread is a weighted measure of how widely the significant path delays are spread:

\displaystyle \sigma_\tau=\sqrt{\frac{\sum_i P_i(\tau_i-\bar{\tau})^2}{\sum_i P_i}}

where P_i is path power and tau_bar is mean delay:

\displaystyle \bar{\tau}=\frac{\sum_i P_i\tau_i}{\sum_i P_i}

Large delay spread can make a narrowband signal fade differently across frequency or make symbols interfere with each other in a wideband waveform.

Coherence Bandwidth

A common early screening approximation is:

\displaystyle B_c\approx\frac{1}{5\sigma_\tau}

If sigma_tau=0.8 microseconds, then:

\displaystyle B_c=\frac{1}{5(0.8\times10^{-6})}=250\ \text{kHz}

A signal much narrower than this may see mostly flat fading. A signal much wider than this can see frequency-selective fading, notches and unequal subcarrier or symbol response. Actual receiver behavior depends on equalization, pilot spacing, coding, interleaving, cyclic prefix, channel estimation and implementation limits.

Motion and Doppler

Multipath fading can change with motion. A Doppler screen is:

\displaystyle f_D=\frac{v}{\lambda}

where v is relative speed and lambda is wavelength. A rough coherence-time screen is:

\displaystyle T_c\approx\frac{0.423}{f_D}

Fast motion or moving reflectors can make channel estimates stale. A receiver that works in a static bench test may need faster pilots, tracking loops, diversity or a more conservative modulation mode in the field.

Engineering Response

Mitigations depend on the failure mode. Antenna relocation, height change, directional antennas, polarization changes, diversity, MIMO, equalization, interleaving, adaptive modulation, lower-order modulation, narrower bandwidth, longer guard interval or route changes may help. Increasing transmit power may not solve a deep frequency notch or excessive delay spread; it can also worsen interference or violate EIRP limits.

Validation Evidence

A defensible multipath review includes received-power distribution, SNR or SINR distribution, EVM, packet-error or BER data, channel impulse response where available, delay-spread evidence, Doppler or mobility assumptions, antenna locations, surrounding reflectors, spectrum occupancy, modulation mode logs, equalizer status and field conditions. For OFDM systems, cyclic-prefix margin and subcarrier EVM are especially useful.

The evidence should separate multipath from other causes. Rain fade, Fresnel obstruction, pointing loss, polarization mismatch, receiver desensitization and interference can all reduce service margin, but they require different corrective actions.

Common Mistakes

Common mistakes include validating only average RSSI, treating multipath as ordinary path loss, ignoring delay spread for wideband signals, assuming an equalizer can fix every channel, using one indoor test location as proof of coverage, and adding fade margin without stating the percentile or service requirement.

The practical rule is to reserve margin for fading, measure the channel under credible field conditions and tie the waveform choice to delay spread, Doppler and validation evidence.

REF

Disciplines