应用信号变化度的线性混合信号盲分离算法

Signal Variability Based on Linear Mixture Blind Source Separation Algorithm

论文首先给出了信号变化度的概念,并证明了信号变化度的一个性质:互相独立的一组源信号的线性混合信号的变化度介于源信号中的最小变化度和最大变化度之间。然后,利用矩阵广义特征值理论,给出了一种基于线性混合信号盲分离算法。该算法计算简单,具有闭解形式;并能分离源信号中既有亚高斯信号又有超高斯信号的情况。仿真结果表明该算法是有效的,并具有很好的分离性能。

In this paper,a measure of signal variability is defined.Given any set of statistically uncorrelated source signals,it is proved that a linear mixture of those signals has the following property：the signal variability of any signal mixture is greater than（or equal to） minimal that of its component source signals,and is less than（or equal to） maximal that of its component source signals.Based on the property,an algorithm for linear blind source separation is proposed by using generalized eigenvalue theory.The proposed algorithm has a closed-form and less computations,And the presented algorithm can separate signal mixtures in which each mixture is a linear combination of source signals with supergaussian,subgaussian,and gaussian probability density functions.Simulation results illustrate the efficiency and the good performance of the algorithm.