摘要
针对盲源分离信号顺序的不确定性以及基于峭度的盲抽取算法运算量大的问题,本文提出一种基于偏度的低空目标声信号盲抽取算法。该算法利用偏度的非对称性,将其作为信号非高斯性的度量,抽取出的信号可以按偏度绝对值的降序排列。与基于峭度的盲抽取算法相比,该算法大大减小了运算量,同时具有较好的分离效果,且对于噪声具有鲁棒性。声音信号及战场低空飞行目标声音信号盲抽取的实验结果,验证了该算法的有效性。
For the uncertainty of the signal sequence of blind source separation and the large computation of the kurtosis-based blind source extraction algorithm,a blind signal extraction algorithm for low altitude acoustic based on skewness is proposed.An algorithm uses the asymmetric of skewness to evaluate the non-Gaussianity of a signal.The signal can be extracted by the absolute value of the skewness in descending order.Compared with the kurtosis-based blind extraction algorithm,the algorithm reduces the computational complexity,has better separation efficiency and is robust for the noise.Computer simulation and experiments on the acoustic signal and low altitude battlefield acoustic signal confirm the effectiveness of the algorithm.
出处
《数据采集与处理》
CSCD
北大核心
2011年第1期69-73,共5页
Journal of Data Acquisition and Processing
基金
国家自然科学基金(60872113)资助项目
关键词
盲抽取
盲分离
偏度
blind source extraction
blind source separation
skewness
