【摘要】 本文介绍了一种高效的盲分离算法———基于独立性的等变自适应分离算法 ,简称EASI算法 (EASI:Equiv ariantAdaptiveSeparationviaIndependence)。为了克服EASI算法依据峭度选择非线性函数的问题 ,引进了其改进算法。将其运用到声学诊断中 ,对声混合信号进行频谱分离 ,得到各声源的特征频谱结构图。EASI及其改进算法经过实验研究 ,证明了该算法的有效性。更多还原

【Abstract】 In the paper an efficient algorithm for blind source separation,called equivariant adaptive separation via independence(EASI) is introduced.Its extended algorithm is also introduced.EASI and its extended algorithm are used in acoustic diagnosis field.With the help of EASI and its extended algorithm as well as the knowledge of fault diagnosis,the mixed signal is seperated to get every source’s spectrum.The simulation and experiment demonstrate the effectiveness of the method proposed.更多还原