Array Signal MP Decomposition and Its Preliminary Applications to DOA Estimation

Abstract

The idea of sparse decomposition is introduced into array signal processing, and a novel approach to DOA estimation is presented in this paper. The approach decomposes the array signal over an over-complete dictionary, the atoms of which are vectors established according to the array geometry. The sparse decomposition is implemented by matching pursuit (MP) in the proposed algorithm. High resolution of DOA estimation can be obtained according to the parameters of the atoms decomposed with MP. The DOA estimation resolution capabilities are shown to be much higher than MUSIC and ESPRIT, especially in the case of less array elements and lower SNR. Furthermore, the performance is not affected by the correlation of the signals to be resolved. Computer simulation confirms its validity.