Delay Spread

The most common statistical measure, calculated as the square root of the second central moment of the power delay profile. It weights multipath components by their relative power, providing a single number that characterizes the effective time dispersion of the channel.

• Calculation: The standard deviation of the delay of multipath components, weighted by their power.
• Typical Values: Indoor office environments: 10-100 ns; Urban macrocellular: 1-10 µs; Hilly terrain: 10-20 µs.
• Significance: Directly determines the maximum achievable symbol rate before equalization becomes mandatory.

The first moment of the power delay profile, representing the average delay of all multipath components relative to the first arriving signal. While simpler to compute than RMS delay spread, it is less informative about the severity of time dispersion.

• Calculation: The power-weighted average of all excess delays in the channel impulse response.
• Limitation: A channel with two strong clusters widely separated in time can have the same mean excess delay as a channel with a continuous, low-level diffuse tail, despite vastly different ISI potential.
• Use Case: Often reported alongside RMS delay spread to provide a complete first-order statistical picture of the channel's temporal structure.

The total time span of the power delay profile, measured from the first arriving component to the last component that exceeds a specified threshold above the noise floor. This defines the temporal boundary of the channel's impulse response.

• Threshold Dependence: Typically defined at -20 dB or -30 dB relative to the strongest peak. The choice of threshold significantly impacts the measured value.
• Coherence Bandwidth Relationship: The maximum excess delay is inversely proportional to the coherence bandwidth. A large τmax implies a small coherence bandwidth, indicating severe frequency selectivity.
• Guard Interval Design: In OFDM systems, the cyclic prefix duration must exceed τmax to completely eliminate ISI.

The foundational measurement from which all delay spread statistics are derived. The PDP plots received power as a function of time delay relative to the first arrival, visualizing the multipath structure of the channel.

• Representation: Typically displayed as a bar chart or continuous function showing distinct multipath clusters and their relative amplitudes.
• Measurement: Obtained by transmitting a wideband probing signal and correlating the received signal with a known replica, effectively capturing the channel impulse response squared.
• Temporal Variability: The PDP is not static; it evolves as scatterers in the environment move. Averaging over time or space is often required to obtain a representative profile.

A channel is considered frequency-selective when the signal bandwidth exceeds the coherence bandwidth of the channel. Delay spread is the direct cause: if the RMS delay spread is large relative to the symbol period, different frequency components of the signal experience uncorrelated fading.

• Rule of Thumb: Frequency-selective fading occurs when στ > 0.1 × Ts, where Ts is the symbol period.
• Consequence: Without equalization, the received signal suffers from ISI, where energy from one symbol smears into adjacent symbols, dramatically increasing the bit error rate.
• Mitigation: Adaptive equalizers, OFDM with cyclic prefix, and rake receivers in CDMA systems are all designed to combat the effects of delay spread.

The statistical measure of the frequency range over which the channel response is considered flat or highly correlated. It is the frequency-domain dual of delay spread, with an inverse relationship.

• 50% Correlation: Bc ≈ 1 / (5στ). Frequencies within this range experience correlated fading.
• 90% Correlation: Bc ≈ 1 / (50στ). A more conservative definition for applications requiring high correlation.
• System Design Impact: If the transmitted signal bandwidth is less than Bc, the channel is flat fading and no equalizer is needed. If greater, the channel is frequency-selective and requires complex receiver processing.

The statistical inverse of delay spread, coherence bandwidth defines the frequency range over which the channel response is considered flat or highly correlated. If signal bandwidth exceeds the coherence bandwidth, the channel becomes frequency-selective, causing intersymbol interference. For a channel with RMS delay spread (\sigma_\tau), the coherence bandwidth (B_c) is approximated as (B_c \approx 1/(5\sigma_\tau)). This relationship is fundamental to OFDM subcarrier spacing design.

The time-domain characterization of a wireless channel's multipath profile, representing received signal power as a function of delay when a perfect impulse is transmitted. The power delay profile (PDP) is derived from the squared magnitude of the CIR. Key parameters extracted include mean excess delay, RMS delay spread, and maximum excess delay. In RF digital twins, the CIR is generated via ray tracing or measured in anechoic chambers.

The Wide-Sense Stationary Uncorrelated Scattering assumption is a foundational simplification in stochastic channel modeling. It states that: (1) the channel's statistical properties are stationary over short time intervals, and (2) scatterers at different path delays are uncorrelated. Under WSSUS, the channel is fully characterized by its scattering function—a 2D power spectral density over delay and Doppler domains. This assumption enables tractable delay spread analysis.

The primary impairment caused by excessive delay spread. When the maximum excess delay exceeds the symbol period, energy from one symbol spills into adjacent symbol intervals, corrupting detection. Mitigation techniques include: adaptive equalization (time-domain), OFDM with cyclic prefix (frequency-domain), and rake receivers (CDMA). The cyclic prefix length in OFDM must exceed the maximum expected delay spread to eliminate ISI entirely.

The PDP plots received signal power against excess delay, visually representing the multipath structure. From the PDP, the RMS delay spread is calculated as the second central moment of the profile. Common PDP shapes include exponential decay (indoor), uniform (reverberation chambers), and clustered (3GPP models). RF digital twins must accurately reproduce these profiles to validate equalizer and channel estimation performance.

While delay spread quantifies time dispersion, Doppler spread quantifies frequency dispersion caused by relative motion. The two are duals: delay spread causes frequency selectivity, while Doppler spread causes time selectivity. Together they define the channel's spread factor. In high-mobility scenarios, the channel is doubly selective, requiring joint delay-Doppler estimation techniques like OTFS modulation for reliable communication.