摘要
将一种基于偏差去除技术的自然梯度算法应用于噪声盲分离中,估计出分离矩阵,并采用一种新的小波阈值函数对已分离出的含噪信号进行去噪处理。通过计算机仿真验证,该算法分离效果良好,盲分离结果具有更高的信噪比。
This paper used a method which was based on bias removal technology of natural gradient algorithm under the noisy circumstances to estimate the demixing matrix. Then applied a new wavelet threshold function to separated signals to remove noise. The simulations verify that this algorithm is robust and improves the SNR of the ultimate results obviously.
出处
《计算机应用研究》
CSCD
北大核心
2010年第8期2886-2888,共3页
Application Research of Computers
基金
湖南省湘潭市科技计划项目(ZJ20071008)
关键词
盲源分离
自然梯度算法
偏差去除技术
小波阈值去噪
blind source separation( BSS)
natural gradient algorithm
bias removal technique
wavelet threshold denoising