The signal subspace decomposition method for extracting harmonic signal

Abstract

The subspace decomposition method for extracting harmonic character is discussed. The theory of the cross-correlation matrix based MUSIC method in Gaussian white noise is analysed. A method of subspace decomposition is proposed for the purpose of suppressing noise interference and increasing the ability of recognizing targets and multiple signal classification, an algorithm of harmonic retrieval, was introduced to extract harmonic frequencies. Then the frequency of harmonics is estimated by setting threshold in MUSIC power spectrum curve. The simulation proves the method is correct and feasible