Carrier Frequency Offset Estimation

The Spectral Correlation Function is the fundamental two-dimensional transform that reveals the correlation between frequency-shifted versions of a signal. For CFO estimation, the SCF exhibits distinct peaks at specific cyclic frequencies that are directly proportional to the carrier offset. By locating these peaks, the offset can be estimated without any pilot tones or training sequences. The SCF is defined as the Fourier transform of the cyclic autocorrelation function and is the primary visualization tool for identifying hidden periodicities in a signal's spectral content.

The cyclic autocorrelation function is the time-domain precursor to the SCF, measuring the correlation between a signal and a frequency-shifted, conjugated version of itself. It is parameterized by the cyclic frequency (α) and the lag parameter (τ). For a signal with a carrier offset, the cyclic autocorrelation exhibits non-zero values at cycle frequencies corresponding to multiples of the symbol rate plus the offset. This function is the mathematical foundation upon which all blind CFO estimators are built, transforming a timing problem into a frequency-domain peak detection task.

The FFT Accumulation Method (FAM) is the most widely implemented computationally efficient algorithm for estimating the spectral correlation function. It operates by channelizing the input signal using a sliding window, computing short-time FFTs, and then correlating frequency-shifted outputs. For CFO estimation, the FAM algorithm allows real-time or near-real-time extraction of cyclostationary features, making it practical for deployment in software-defined radios. The algorithm trades off cycle frequency resolution against spectral frequency resolution through its channelizer design parameters.

The cyclic frequency, universally denoted by the Greek letter alpha (α), is the independent variable that quantifies the periodicity of a signal's statistical moments. In the context of CFO estimation, the carrier offset manifests as a shift in the expected cyclic frequencies of the modulation scheme. For example, a BPSK signal has cyclic features at α = k/T_sym ± 2f_c, where f_c is the carrier frequency. By measuring the deviation of these peaks from their nominal positions, the CFO is directly estimated. Alpha is the axis along which modulation-specific signatures are read.

The cyclic periodogram is the simplest, though statistically inconsistent, estimator of the spectral correlation function. It is computed directly from the product of two frequency-shifted, finite-time Fourier transforms of the observed signal. While its variance does not decay to zero with increasing observation time, it serves as the basic building block for more sophisticated estimators. Understanding the cyclic periodogram is essential for grasping the time-frequency resolution trade-off inherent in all cyclostationary CFO estimation techniques and for diagnosing spectral leakage artifacts.

Second-order cyclostationarity is the specific property exploited for blind CFO estimation, defined by a periodic autocorrelation function. Most man-made communication signals exhibit this property due to coupling between their symbol rate and carrier frequency. The CFO estimation process relies entirely on analyzing these second-order statistics, as the carrier offset directly modulates the periodicity of the autocorrelation. This property is what distinguishes a cyclostationary approach from simpler stationary methods that cannot separate the offset from the modulation without a known reference.