A fast and accurate algorithm for estimating frequency of a sine wave under low-signal-noise-ratio background

Abstract

A new method of estimating frequency of a sine wave with the existence of serious Gaussian noise is proposed in this paper, in which the accurately-estimated frequency is obtained by two procedures as coarse search and fine search based on DFT and interpolation. The method is used to estimate the beat frequency of a LFMCW radar liquid-level gauge and its distance-resolution is up to 0.3 mm under noise background as serious as -32 db of signal-noise-ratio. The method, proved by simulation results, is of high accuracy and capable of being used as an on-line way of frequency estimation.