贝叶斯超分辨方法探讨

Discussion on a Bayesian High Resolution Approach

根据贝叶斯最大后验估计原理 ,研究了波达方向及频率估计的超分辨方法 ,提出了多目标波达方向及频率估计的快速算法 ;对等间距阵列模型予以推广 ,得到非等间距阵列的贝叶斯超分辨模型 ;对双目标和三个目标进行了仿真。仿真结果表明 ,该方法表现出较准确、稳定的分辨能力 ,分辨能力达 1/ 10功率波束宽度。

By using a Bayesian maximum likelihood(BML) technique, a high resolution algorithm is discussed for direction of arrival(DOA) and frequency estimation. A fast algorithm is presented for multi targets and it is extended to an array antenna which elements are not equally spaced. It has been shown, by the simulations of two and three sources, that the Bayesian technique has significant advantages over MUSIC and other eigen decomposition methods in the cases of low signal to noise ratios and closely spaced sources (e. g. 1/10 3dB beamwidth).