Rayleigh Fading

A statistical model for propagation environments where a dominant line-of-sight (LOS) signal component exists alongside scattered multipath components. The K-factor quantifies the power ratio between the dominant LOS path and the scattered paths. When the K-factor approaches zero, Rician fading collapses to Rayleigh fading, making Rayleigh a special case of the Rician model. Modulation classifiers must distinguish between these regimes, as the signal envelope distribution shifts from Rayleigh to Rician, affecting feature extraction.

The improvement in link reliability achieved by transmitting redundant copies of the signal over independently fading spatial paths. In a Rayleigh fading environment, the probability of all paths experiencing a deep fade simultaneously decreases exponentially with the number of diversity branches. Techniques include:

• Space diversity: Multiple antennas spaced sufficiently apart
• Frequency diversity: Transmitting across different carrier frequencies
• Time diversity: Repeating transmissions separated by the coherence time For modulation recognition, diversity combining can stabilize the received signal before classification.

The known channel properties of a communication link, including scattering, fading, and power decay, used by a transmitter to adapt its signal to current propagation conditions. In Rayleigh fading channels, CSI captures the complex Gaussian channel coefficients. Accurate CSI estimation is critical for:

• Coherent detection in MIMO receivers
• Precoding matrix selection for beamforming
• Modulation classification algorithms that use channel-compensated features Imperfect CSI due to estimation errors in fast Rayleigh fading can severely degrade classifier performance.

The statistical dependence between antenna elements caused by insufficient spacing or a sparse scattering environment. In ideal Rayleigh fading, antenna elements are assumed uncorrelated, but in practice, correlation degrades MIMO channel rank and capacity. Key factors:

• Antenna spacing: Typically requires λ/2 or greater
• Angular spread: Narrow spread increases correlation
• Scatterer richness: Fewer scatterers reduce spatial diversity Modulation classifiers trained on uncorrelated Rayleigh channels may fail when deployed in highly correlated environments, necessitating robust feature engineering.

The process of characterizing the propagation channel's impulse response using known reference signals or pilot symbols. In Rayleigh fading channels, estimators must track the rapidly varying complex Gaussian coefficients. Common approaches:

• Least Squares (LS): Simple but noise-sensitive
• Minimum Mean Square Error (MMSE): Optimal but requires channel statistics
• Deep learning-based: Neural networks trained on Rayleigh fading realizations Accurate channel estimation enables coherent demodulation and provides compensated IQ samples for downstream modulation classification.

A metric describing the sensitivity of a MIMO channel matrix to inversion, defined as the ratio of the largest to smallest singular value. In Rayleigh fading channels:

• Well-conditioned channels (low condition number): Support high spatial multiplexing gains
• Ill-conditioned channels (high condition number): Indicate spatial correlation or keyhole effects For modulation recognition, a high condition number amplifies noise during zero-forcing equalization, distorting the signal constellation and confusing classifiers. MMSE receivers mitigate this by balancing interference suppression with noise enhancement.