中文摘要：

为了解决线性传声器阵列采集的高压水射流反射声解耦、重构和定位问题,以传统独立成分分析（ICA）理论的旋转因子乘积法为基础,构建了一种迭代式ICA方法,介绍了其算法原理.设计了线性阵列排列的多通道水射流探测与反射声采集系统进行实验,用多路仿真声音信号和传声器阵列实际采集的反射声信号进行算法验证.结果表明：构建的迭代式ICA方法能够精确分解、重构线性传声器阵列采集的高压水射流靶物反射声信号,而且可以消除环境噪声及不同声源干扰信号的影响,重构信号与源标准信号的顺序一致,保留源标准信号的大部分特征信息,可以用于后续的高压水射流靶物分类识别与定位.

英文摘要：

In order to solve the decoupling,reconstruction and location problems of high pressure water jet reflected sound signal acquired by linear microphone array,a new iterative indepentdent component analysis（ICA）method was built based on the rotation product factor method of traditional ICA theory.Its algorithm principle was introduced in detail.A linear arranged multi-channel water-jet detecting system and data acquisition system was designed to do the sound signal acquisition experiments.The multi-channel simulation sound signals and actual reflected sound signals were used to verify the ICA method.The verification results show that the iterative ICA method can decompose and reconstruct the reflected sound signals of high-pressure water jet targets acquired by linear microphone array accurately.The interference influence from the environmental noise and other sound sources can be eliminated.The order of reconstructed reflected sound signal is equal to standard signal and most characteristic information of standard signals is retained.So the iterative ICA method can be used for targets classification and location of high pressure water jet.