声矢量传感器阵列在非理想传输条件下的声源定位研究

Acoustic Vector Sensor (AVS) has comprehensive applications in the area of source localization due to its advantages of direction sensitivity and noise resistance. Nowadays, most of source localization algorithms with AVS array assume the ideal conditions, such as far-field source, open space, narrowband signals, no failed sensor, and etc. However, as the developing of the informational society and the rising of internet of things, many application environments turn to be more compact, professional, and complicated. The thorough study of source localization under the non-ideal conditions has great significance. In this proposal, by exploring the distributed signal processing feature of the AVS array, we investigate a near-field source localization method, which is capable of locating the wideband sources under the adverse environments with attenuation, reverberations, multipaths and even sensor failure. The basic idea of the method is described as follows: First, the DOA estimates and the estimation errors from all AVS nodes are combined to determine a zone in which the source is likely to locate. Then in this zone, the steered response power is used to combat the reverberation and locate the wideband coherent sources. Lastly, the weights of the AVS nodes are re-distributed to distinguish the true locations of the sources. Thus, the sources can be located with reliability. With this method, we will also optimize the algorithm, reduce its computation load, and at last assess its performance of localization.

声矢量传感器（AVS）以其优越的空间指向性和抗噪声性能在声源定位领域有着广泛的应用。目前，许多声矢量阵的定位算法大都基于远场声源、开放空间、窄带信号、无失效阵元等理想化的假设条件。然而，随着社会信息化程度的加深和物联网的兴起，许多应用环境正朝着局域化、专门化、和复杂化的方向发展，对非理想条件下的声源定位算法进行深入研究具有重要的理论价值和现实意义。通过充分利用AVS阵元的分布式信号处理特性，本项目研究了一种近场声源定位方法，可以在混响、多径甚至阵元失效等不利条件下，实现对近场宽带声源的三维定位。其基本思路是：首先联合所有AVS阵元对波达方向的估计和误差来确定声源的预估判定区域，再利用基于响应功率的抗混响算法对该区域内的宽带声源进行定位，最后通过阵元权重再分配等方法对真实声源进行识别，从而达到声源可靠定位的目的。利用该方法，我们还将对算法进行优化，减小其运算量，并对其定位性能进行评估。

声矢量传感器（AVS）以其优越的空间指向性和抗噪声性能在声源定位领域有着广泛的应用。目前，许多声矢量阵的定位算法大都基于远场声源、开放空间、窄带信号、无失效阵元等理想化的假设条件。然而，随着社会信息化程度的加深和物联网的兴起，许多应用环境正朝着局域化、专门化、和复杂化的方向发展，对非理想条件下的声源定位算法进行深入研究具有重要的理论价值和现实意义。通过充分利用AVS阵元的分布式信号处理特性，本项目研究了一种近场声源定位方法，可以在混响、多径甚至阵元失效等不利条件下，实现对近场宽带声源的三维定位。其基本思路是：首先联合所有AVS阵元对波达方向的估计和误差来确定声源的预估判定区域，再利用基于响应功率的抗混响算法对该区域内的宽带声源进行定位，最后通过阵元权重再分配等方法对真实声源进行识别，从而达到声源可靠定位的目的。利用该方法，我们还将对算法进行优化，减小其运算量，并对其定位性能进行评估。