Power Spectrum Estimation Method Based on Matlab

Abstract

Power spectrum estimation is one of the important research contents of digital signal processing. Power spectrum estimation is divided into classic power spectrum estimation and modern power spectrum estimation. Modern power spectrum estimation is proposed for the shortcomings of classical power spectrum estimation. The principles of the periodogram method, the improved welch method and the AR model method in the classic power spectrum estimation, Matlab simulations were carried out, and their characteristics were analyzed and compared. It was found that the Burg method of the AR parameter model is better. The classic power spectrum estimation has large variance and low spectral resolution, but modern power spectrum estimation is not affected by the window function, so it has higher spectral resolution and smooth spectral curve.