High-accuracy frequency estimation algorithm based on autocorrelation and phase difference method

In order to improve the frequency estimation accuracy of sinusoidal signal in white Gaussian noise, the correlation length m of the algorithm based on the autocorrelation operation was derived, and the equation representing the relationship of m and signal-to-noise ratio (SNR) was acquired. Given the data length N, the optimum value of m was N/3 when the SNR was large and it was N/2 when the SNR was small. The performance of segmented-FFT phase difference method and autocorrelation method were analyzed, and a new frequency estimation algorithm was proposed. The Monte Carlo simulation results verify that the new algorithm combines the advantages of the two methods previously mentioned, and avoid the disadvantages of them. It can fetch the highly accurate frequency of a signal in a wide range of SNR, and the computational cost is small.