采用条件波数谱密度函数的宽带高分辨方位谱估计算法

针对宽带高分辨方位估计存在方位估计偏差大、算法复杂度高等问题,提出了一种基于条件波数谱密度(Conditional Wavenumber Spectral Density based,CWSD-based)的宽带高分辨方位谱估计算法.该算法利用条件波数谱密度将阵列信号转换到频率-波数空间,宽带信号能量在该空间的坐标呈现与入射角相关的线性分布,通过借鉴直线检测原理,实现邻近目标的高分辨方位估计,且无需预估角度和信源数等信息。仿真结果表明,该算法理论分辨率与处理最高频率成反比,估计均方误差约为0.1°,对阵形畸变鲁棒,运算效率高。海上试验数据表明,本文方法在方位分辨率、弱目标检测、非目标向噪声抑制、稳健性等方面都优于宽带常规波束形成和最小方差无畸变算法,在实际海洋中可实现超低旁瓣高分辨波达方向估计。 

Broad-band high-resolution direction-of-arrival estimation of a conditional wavenumber-spectral-density based approach

A broad-band high-resolution Direction-of-Arrival(DOA) estimation algorithm named the Conditional Wavenumber Spectral Density based(CWSD-based) is proposed in this paper aimed at the problem of large error and high complexity of wide-band.The super-resolution DOA estimation,minimal main-lobe width,ultra-low side-lobe and high precision target resolution performance can be obtained after transforming the array signals to f-k space and using the characteristic of wideband signal power distributed as a line on the Conditional Wavenumber Spectral Density.Moreover,there are no DOA pre-estimates of the incident signals and the number of sources to be required in this algorithm.Computer simulations indicate that the theoretical resolution of the algorithm is inversely proportional to the highest processing frequency,the estimated mean square error is about 0.1 degree,and also is robust to array elements distortion.Meanwhile,it achieves lower time complexity.The proposed algorithm,whose performance is superior to CBF and MVDR in the aspects of resolution,weak target detection,non-target direction noise suppression and robustness etc.,can achieve ultra-low sidelobe high-resolution DOA in real marine environment.