Pilot-Induced Cyclostationarity

Pilot-induced cyclostationarity is a man-made statistical signature created by multiplexing a known, deterministic sequence (pilots) with random data. Unlike natural cyclostationarity arising from modulation or symbol rate, this periodicity is intentionally engineered into the waveform.

• The pilot repetition rate defines a precise cyclic frequency (alpha)
• This creates strong, predictable correlation peaks in the Cyclic Autocorrelation Function (CAF)
• The resulting signature is independent of the random data payload, providing a stable reference for receivers

The same pilot structure serves two critical functions simultaneously, making it a highly efficient physical layer design choice.

• Channel Estimation: The receiver correlates the known pilot sequence with the received signal to estimate the channel impulse response for equalization
• Emitter Identification: The unique pilot pattern, timing offset, and power allocation create a device-specific cyclostationary fingerprint that can be extracted for authentication
• This dual-use eliminates the need for separate overhead for security and channel sounding

The specific arrangement of pilots in the time-frequency grid directly determines the cyclostationary signature's structure and robustness.

• Pilot Density: Higher pilot density increases cyclic feature strength but reduces spectral efficiency
• Pilot Spacing: The spacing in time and frequency defines the cyclic frequency (alpha) and spectral frequency (f) coordinates of correlation peaks in the SCF
• Pilot Sequence: Orthogonal sequences (Zadoff-Chu, Gold codes) minimize cross-correlation between devices, enabling multi-user separation via Cyclic MUSIC or Cyclic DOA Estimation

A primary advantage of pilot-induced cyclostationarity for identification is its inherent immunity to stationary Gaussian noise and wide-sense stationary interference.

• Noise lacks periodicity, so its cyclic autocorrelation is zero for non-zero cyclic frequencies
• The Cyclic Cumulant of the pilot structure is non-zero, while Gaussian noise cumulants vanish above second order
• This allows reliable extraction of the pilot signature even at negative signal-to-noise ratios (SNR), critical for long-range or low-power device authentication

Pilot-induced cyclostationarity is fundamentally different from the cyclostationarity generated by the random data payload or modulation format.

• Data-Induced: Cyclic frequencies are tied to symbol rate, carrier offset, or guard interval — common across all devices using the same waveform standard
• Pilot-Induced: Cyclic frequencies and correlation patterns are configurable per device or per transmission, enabling unique identification
• This separation allows a receiver to use Cyclic Feature Detection to distinguish between standard-compliant signals and specific, authenticated emitters

Orthogonal Frequency-Division Multiplexing (OFDM) systems, such as 4G LTE and 5G NR, heavily rely on pilot-induced cyclostationarity through reference signals embedded in the resource grid.

• Cell-Specific Reference Signals (CRS) in LTE create a cyclostationary signature at the slot rate
• Demodulation Reference Signals (DMRS) in 5G NR are user-specific, enabling per-device cyclostationary fingerprinting
• The OFDM Cyclic Prefix adds an additional layer of cyclostationarity, but the pilot structure provides a more controllable and unique identifier