The main studied topic of spatial spectrum estimation is the direction of arrival (DOA) of spatial signals. The super-resolution algorithm based on modern spectrum analytical theory intends to improve DOA estimation quality in the presence of dense signals. The detailed content of this paper is arranged as follow:Firstly, the classical super-resolution algorithms, mainly including MUSIC algorithm, ESPRIT algorithm are introduced.Secondly, based on the theoretical analysis of the effect of element's gain and phase uncertainty on the performance of the MUSIC algorithm in DOA estimation, a universal technique for calibration which is based on ML estimation and Toeplitzation is proposed. Computer simulation shows that it can not only alleviate its sensitivity to the array uncertainty, but further decrease its threshold of SNR.Finally, based on the theoretical analysis of MMUSIC algorithm, a new modified MUSIC algorithm is presented. It can estimate the DOA of coherent signals and adjacent signals with small SNR effectively. Computer simulation confirms the new method availability. |