论文部分内容阅读
基于被动合成孔径原理,提出了一种具有高稳健性的运动声源高分辨聚焦定位识别方法.该方法采用综合优化手段,通过矢量最大似然聚焦定位算法生成虚拟阵列坐标及数据矩阵,进而利用基于最差性能优化的稀疏虚拟阵列聚焦算法,获取稳健的高分辨定位识别效果.理论及仿真研究表明,该方法对于非匀速运动以及与基阵存在运动倾角的复杂情况具有较强的适用性,聚焦空间谱表现出更大的动态范围、更为尖锐的聚焦峰尺度以及更强的背景噪声起伏压制能力.湖上试验进一步验证,在高分辨最小方差信号无畸变响应法(MVDR)聚焦算法动态范围仅为3.5dB的相同试验条件下,该方法的动态范围可优于50dB,具有更高的稳健性和更好的定位识别效果.
Based on the principle of passive synthetic aperture, a high-resolution robust localization method for motion sound sources is proposed. This method uses a synthetic optimization method to generate virtual array coordinates and data matrix by using vector maximum likelihood focal positioning algorithm. Based on the sparse virtual array focusing algorithm with the worst performance optimization, a robust high-resolution localization recognition effect is obtained.The theoretical and simulation studies show that the proposed method has strong applicability to the non-uniform motion and the complex situation of the dip angle with the matrix, The focused spatial spectrum exhibits greater dynamic range, sharper focus peak size, and stronger background noise suppression capability. The Lake test further validates the MVDR algorithm for high-resolution minimum variance signal focusing algorithms. Dynamic Range Under the same experimental condition of only 3.5dB, the dynamic range of this method can be better than 50dB, which has higher robustness and better positioning and recognition.