Signal frequency estimation based on modified window scaling constant with maximum deviation criterion

Weighting windows may have effect on the estimation performance of signal frequency, thus modified window scaling WSC (WSC) should be used for revising the error. Frequency estimation accuracy varies with different position between two samples in spectrum when the frequency of a tone is contaminated with noise. Only considering bias and root mean square error (RMSE) at each position cannot give a global optimized WSC. Maximum deviation criterion (MDC) is proposed for selecting the modified WSC in this paper. The proposed method presents a performance quantitative criterion considering bias and RMSE at different sampling locations with 3σ criterion which fits for signal frequency estimation. Performance of MDC is analyzed by Monte Carlo simulation. The result shows that using the modified WSC based on MDC can increase the whole performance of signal frequency estimation.