Method for estimating the central frequency of phase-coded radar signals

The authors analyse radar carrier frequency estimation errors in digital electronic support measures receivers. They assume that the received signal is phase-coded, real-valued and embedded in white Gaussian noise. The corresponding Cramer–Rao lower bound (CRLB) for the frequency estimation is derived and formulated. The performance degradation can be significant when just a single traditional fast Fourier transform (FFT) algorithm used to estimate the central frequency of a phase-coded signal, where the spectrum is wide and distributes across many frequency bins. In this study, they assume that the FFT magnitude response of the received signal can be approximated by a raised Gaussian-shaped function. The corresponding Gaussian function fitting (GFF) estimator is proposed as a peak frequency estimation technique. Typical FFT-based frequency estimation process is simulated and the performance of the GFF estimator is compared with that of the most commonly used peak estimators: namely, with the maximum point (MAX) and the spline interpolation (INT). GFF estimator performance is also compared with a K-means-based peak estimator. With extensive set of simulations for different code lengths, they demonstrate that the proposed technique significantly reduces carrier (central) frequency estimation errors while illustrating CRLB within the figure.