Fading Simulation

A statistical model for the stochastic fluctuation of a signal envelope in a propagation environment with no dominant line-of-sight path. The received signal magnitude follows a Rayleigh distribution, representing the worst-case multipath scenario where numerous reflected and scattered components arrive at the receiver with random phases.

• Models dense urban environments with no direct path
• Envelope follows a Rayleigh probability density function
• Phase is uniformly distributed over [0, 2π]
• Critical for testing receiver sensitivity in NLOS conditions

A statistical model that includes a dominant line-of-sight component along with scattered multipath components. The Rician K-factor quantifies the power ratio between the direct path and the scattered paths, with higher K-factors indicating a stronger LOS component.

• K-factor = 0 reduces to Rayleigh fading
• K-factor > 10 approximates AWGN with slight fading
• Models suburban, rural, and indoor environments with partial LOS
• Essential for testing beamforming and directional antenna systems

The augmentation of RF signals with frequency offsets that mimic the relative motion between a transmitter and receiver. Doppler spread causes spectral broadening and time-selective fading, critical for training models deployed in high-mobility environments such as vehicular or aerial communications.

• Maximum Doppler shift: f_d = (v/c) × f_c
• Coherence time inversely proportional to Doppler spread
• Fast fading occurs when symbol duration exceeds coherence time
• Clarke's model generates realistic Doppler spectra for isotropic scattering

A characterization of a multipath channel that describes the received signal power as a function of time delay. The PDP defines key parameters including mean excess delay, RMS delay spread, and maximum excess delay, which determine the frequency selectivity of the channel.

• RMS delay spread determines coherence bandwidth
• Coherence bandwidth ≈ 1/(5 × RMS delay spread)
• Flat fading when signal bandwidth < coherence bandwidth
• Frequency-selective fading when signal bandwidth > coherence bandwidth
• Standardized PDPs: ITU Pedestrian A/B, Vehicular A/B, EPA, EVA, ETU

The algorithmic process of convolving a transmitted RF waveform with a time-varying channel impulse response to produce a faded output signal. Tapped-delay line models implement this by summing delayed and attenuated copies of the input, each multiplied by independent complex Gaussian fading processes.

• Each tap represents a discrete multipath cluster
• Tap amplitudes follow the specified power delay profile
• Tap phases evolve according to the Doppler spectrum
• Sum-of-sinusoids and filtered Gaussian noise are common generation methods
• Enables repeatable, deterministic channel conditions for model validation

The performance discrepancy observed when a model trained on synthetic faded RF data is deployed in a live over-the-air environment. This gap arises from unmodeled physical imperfections including hardware non-linearities, antenna coupling effects, and non-stationary channel behavior that statistical models cannot fully capture.

• Statistical models assume wide-sense stationary uncorrelated scattering (WSSUS)
• Real channels exhibit non-WSSUS behavior in dynamic environments
• Domain randomization during simulation helps bridge the gap
• RF digital twins with ray-tracing reduce sim-to-real mismatch
• Adversarial domain adaptation aligns simulated and real feature distributions

A statistical model for the stochastic fluctuation of a signal envelope in a propagation environment with no dominant line-of-sight (NLOS) path. It assumes the received signal is the sum of many scattered, reflected, and diffracted components, resulting in a complex Gaussian process. The envelope follows a Rayleigh distribution, and deep fades occur frequently, making it the standard model for heavily built-up urban or indoor wireless channels. Simulating Rayleigh fading is critical for stress-testing receiver algorithms against severe amplitude drops.

A statistical model for channels where a dominant line-of-sight (LOS) component exists alongside scattered multipath components. The Rician K-factor quantifies the power ratio between the dominant LOS path and the scattered paths:

• K=0: Reduces to Rayleigh fading (no LOS).
• K=∞: No fading, pure LOS. Simulating Rician fading is essential for modeling suburban, rural, or satellite communication links where a direct path is typically present but still subject to constructive and destructive interference.

A characterization of a multipath channel that describes the received signal power as a function of time delay. It specifies the relative power and arrival times of discrete multipath components (taps). Common standardized PDP models include:

• ITU Vehicular A/B: For mobile urban environments.
• EPA (Extended Pedestrian A): For low-mobility scenarios.
• TDL (Tapped Delay Line): Customizable profiles for arbitrary environments. The PDP is a primary input parameter for channel simulation, defining the temporal dispersion that causes inter-symbol interference (ISI).

The augmentation of RF signals with frequency offsets that mimic the relative motion between a transmitter and receiver. The Doppler spread, characterized by the maximum Doppler frequency f_d = v/λ, causes spectral broadening and time-selective fading. Simulation parameters include:

• Jakes' Model: A classical sum-of-sinusoids method for generating Rayleigh fading with a specified Doppler spectrum.
• Flat vs. Shaped PSD: Defines the power spectral density shape (e.g., classic bathtub for isotropic scattering). Critical for training models deployed in high-mobility environments like vehicular or high-speed rail communications.

The time-domain representation of a wireless channel, modeled as a linear time-variant filter. The CIR is the fundamental output of a fading simulator, convolved with the clean transmitted signal to produce the distorted received waveform. Key attributes include:

• Delay Spread: The time difference between the first and last significant multipath component.
• Coherence Bandwidth: The frequency range over which the channel response is correlated (≈ 1/delay spread).
• Coherence Time: The time duration over which the channel remains approximately constant (≈ 1/Doppler spread). Generating realistic, time-evolving CIRs is the core task of fading simulation for data augmentation.

The algorithmic modeling of physical propagation effects to augment clean RF signals with realistic environmental distortions. Beyond fading, a complete simulation pipeline includes:

• Additive White Gaussian Noise (AWGN): Models thermal noise floor.
• Phase Noise: Models oscillator instability.
• Carrier Frequency Offset (CFO): Models mismatch between transmitter and receiver local oscillators.
• IQ Imbalance: Models gain and phase mismatches in quadrature down-conversion. Combining these impairments with fading creates high-fidelity synthetic data that bridges the sim-to-real gap.