将阵元域高分辨加权子空间拟合算法（wsF算法）推广到了波束域中，利用声系统接收基阵波束输出信息估计目标方位。由于波束输出可以看作是，一个阵元数为波束个数的虚拟阵的阵元输出，而在大多数的水声系统中，参与确定某个来向的信号方位的波束数通常远小于阵元数，从而算法的运算量得以大大降低。针对某水声系统接收基阵波束输出所作的计算机仿真结果表明，波束域wSF算法应用到实际的水声系统时，可以保持其在阵元域中估计目标方位的优越性能。

A wide-band beamforming scheme for uniform circular array is proposed to improve the detection performance for weak passive signals in this paper. In our processing scheme, the wide-band signals sampled from the underwater acoustic array are firstly decomposed into different narrow-band signals using FFT, and then the narrow-band high gain beamforming process is performed for each subband of the FFT outputs. The subband beamformer outputs are squared nd summed using the Eckart weighting coefficients. The length of FFT is selected to make every subband satisfy the condition of narrow-band assumption. The fifty percent of overlapping ratio and-30 dB sidelobe level chebyshev window are used for FFT processing to obtain better system performance. This approach was employed to process some wide-band experimental data recorded from a 24-element circular array at a lake in China. Compared to the conventional beamforming, enhanced detection performance is obtained without improving the complexity of realization.

提出了一种任意结构阵列宽带恒定束宽波束形成的新方法。该方法把任意结构阵列的响应向量表示成以Bessel函数为核函数的级数之和，并将高阶项截断，然后把频带内各个频率点上的阵列响应向量转化到参考频率上，从而得到恒定束宽波束形成向量，使得各个频率上的波束和参考波束相同。同时，本文还给出了一种基于自适应处理的参考波束优化方法。结合均匀分布离散圆阵所作的计算机仿真验证了本文所提方法的有效性。

本文内容分理论研究和实验研究两部分。首先，对高分辨目标方位估计算法IMP提出了两种改善其收敛速度的方法。利用这两种改进方法，可将原算法的收敛速度提高一倍。计算机仿真验证了改进方法的有效性。其次，针对该算法进行了水池实验研究，实验分析结果显示，IMP算法能够有效地分辨和估计非等功率的两个相干信号源。

To overcome the azimuth sampling constraints imposed in synthetic aperture sonar (SAS), many SAS systems employ a single transmitter and multiple-receiver configuration, which can be regarded as a system containing many bistatic transmit-receive pairs. In traditional reconstruction, the multiple receiver system data is often preprocessed into an equivalent single receiver (monostatic) data by using the phase center approximation (PCA) and along track interpolation, and then followed by standard image reconstruction algorithms. But a blurring will occur at the edge of the target area as the length of the receiver array increasing. In this paper, a novel approach is proposed to overcome the blurring problem caused by PCA based on the bistatic model. In this method, an image is firstly reconstructed with the data of each hydrophone independently, and then all the images are coherently added to give the high quality imagery. Simulation results show the feasibility and availability by applying this algorithm.