Improving the accuracy of the frequency measurement by the spectral method.

Improving the accuracy of noise periodic signal frequency measurement is a crucial job in many applications of signal processing such as radar, sonar, speech, radio communication, control, seismic science, etc. For signal analysis, the Fast Fourier Transform (FFT) is frequently employed, but has the problem of spectral leakage. To reduce the impact of this problem, time windows are used. The precision of frequency estimation can be raised by using an appropriate window and an accurate frequency correction coefficient. The interpolation of three spectral lines forms the foundation of a novel frequency estimation approach that is suggested. The feasibility of the proposed algorithm was compared with the Macleod and Voglewede algorithms. As a result of the simulation, it was determined that the suggested technique has lower frequency estimate errors. When employing the Blackman window, the proposed algorithm's highest frequency estimate error was 0.001, i.e., it decreased by 5 and 3 times compared with the error of 0.005, 0.003 for the MacLeod and Voglewede algorithms, respectively, with a sample size of 512 samples. [ABSTRACT FROM AUTHOR]