Subarrays Combined Estimation of DOA Based on Covariance Matrix Pretreatment

Xueli Sheng,Ting Liu,Chaoran Yang,Mengfei Mu,Longxiang Guo
DOI: https://doi.org/10.1109/ieeeconf38699.2020.9389024
2020-01-01
Abstract:In the area of ultra-distant passive detection of abyssal target, complicated noise environment is an essential influence factor of long linear array detection performance, which contains strong noise and non-uniform noise. Apparently, little conventional beamforming (CBF) methods can be applicable. Therefore, a novel subarrays combined estimation of the Direction of Arrival (DOA) technique for broadband signal is proposed, which is based on preprocessed covariance matrix. Covariance matrix pretreatment technique utilizes the difference of energy among matrix elements to reduce the effect of the strong and non-uniform noise. Based on the stability of DOA estimation results of subarrays, weighting fusion method is presented to repress the non-uniform noise further. The simulation proves that the proposed algorithm can be appropriate for target detection with the elements spaced at larger than half wavelength, non-uniform noise and low SNR. This technique can efficiently decrease the background level of spatial spectrum, make beamwidth narrow and improve the accuracy of DOA even the signal energy is smaller than 125 dB. In other words, the SNR of the first element is lower than -35 dB while the last one's SNR is about 5 dB with the non-uniform noise.
What problem does this paper attempt to address?