데이터 기초의 공분산 행렬로 구성된 EV 방법으로부터 다중 정현파의 주파수 추정에 관한 통계적 분석

Statistical Analysis on Frequency Estimation of Multiple Sinusoids from EV with a Data based Covariance Matrix

A Data-based Covariance Matrix(DCM) is introduced in the Eigenvector(EV) method, among subspace methods of estimating multiple sinusoidal frequencies from finite white noisy measurements. It is shown that the EV with the DCM can obtain the true. frequencies from finite noiseless data Some asymptotic results and further improvement on the DCM are also presented mathematically. Monte-carlo simulations are statistically conducted from the view-points of means and standard deviations in the EV's of DCM and Conventional Covariance Matrix(CCM). Simulations show a great promise for using the DCM, particularly for the cases of short data records, closely spaced frequencies and high signal-to-noise ratios.