中文摘要：

盲抽取算法是自信号分离领域的热点研究方向。论文针对现有盲抽取算法运算量较大的不足，提出将度量信号非对称性的偏度作为代价函数，大大减小了运算量，同时解决了抽取信号的排序问题。进一步，根据感兴趣信号的偏度所在的区间范围，算法将SUMT外点法和基于偏度的方法相结合，提出基于偏度先验信息的代价函数，实现了对感兴趣信号的旨抽取。声音信号盲抽取的实验结果验证了本文算法的有效性。

英文摘要：

Blind signal extraction algorithm is a hot research field of blind signal separation. For the large computation of present algorithms, this paper uses the asymmetric of skewness as the cost function to evaluate the non-Gaussianity of a signal. The algorithm reduces the computational complexity greatly, and sloves the problem of order uncertainty. Further, according to the skewness interval of the interested signal, the algorithm combines SUMT-exterior penalty method and skewness-based algorithm, proposes a novel cost function based on priori information of skewness, and achieves the blind extraction of the interested signal. Computer simulation and experiments on acoustic signal confirm the effectiveness of the algorithm.