二阶锥规划在噪声源稳健定位识别中的应用

A robust localization and identification method of noise sources using second-order cone programming

针对噪声源高分辨定位识别方法在失配条件下性能严重退化的问题,提出了稳健矢量阵MVDR聚焦波束形成方法.该方法在已有高分辨聚焦算法的基础之上,通过对矢量阵近场权向量的模强加不等式约束,进而利用二阶锥规划求解最优权及空间谱,明显改善了矢量阵MVDR聚焦处理器的稳健性.优化后的高稳健性聚焦处理器在失配条件下可以获得更大的动态范围、更尖锐的聚焦峰尺度以及更强的背景噪声级抑制能力,可以满足噪声源定位识别对于稳健性的迫切需求.通过理论建模、仿真及湖试试验研究,综合验证了该算方法在失配条件下的有效性,为该方法的实际工程应用提供了理论及试验支持.

In order to improve the robustness of the noise source high-resolution localization and identification meth- od, a new robust minimum variance distortionless response （MVDR） focused beamforming method based on vector sensor array was proposed in this paper. On the basis of previous MVDR focused beamforming, this new method improves its robustness by imposing the inequality norm constraint on the weight vector, and solves the optimal weight vector and spacial spectrum by second-order cone programming. Because this method is robust against the array steering vector mismatch, it can achieve a greater dynamic range, ground sharpe noise level. The validity of this method under mismatched conditions can r focused peak, and lower back- be verified by theoretical model- ing, simulation, and experimental research, and the results provide the theoretical and experimental support for en- gineering applications.