【摘要】 提出一种基于似P范数特征分解的高分辨率声源定位识别方法,该方法在子空间类算法原理基础上,利用特征分解得到子空间响应函数向量,通过预设声源类型建立各子空间声源向量重构模型,进而利用似P范数稀疏性约束条件求解最优解,获取高分辨率声源定位识别效果。理论及仿真研究表明,与其它常规算法相比,该方法不仅能真实反映声源位置信息,而且能反映不同声源能量分布的绝对大小,对多种类型声源具有高精度,高分辨率定位识别效果,适用性强。通过对影响定位性能参数的仿真分析,给出了合理的选取范围。水池试验进一步验证了该方法具有良好的工程应用前景。更多还原

【Abstract】 A high-resolution method for sound sources locating and identification based on LP norm eigenvalue decomposition was proposed here.According to the principle of the signal subspace method,the subspace response function vectors were obtained by performing eigen-decomposition of a cross spectral matrix.The sound source vector reconstruction model of each feature subspace was established via pre-defined sound source types for the reference solutions.Then,by utilizing the sparse constraint condition of LP norm,the optimal solution was solved,the highresolution locating and identification of sound sources was achieved.The theoretical study and simulations showed that the proposed method can be used not only to obtain the sound source locating results but also to reflect the absolute energy contribution of each sound source compared with other existing ordinary algorithms.The parameters affecting the locating performance of sound sources were also reasonably chosen with numerical simulations.The pool tests verified the effectiveness of the method,it was shown that the proposed method has a good prospect for engineering applications.更多还原