基于独立分量分析的混沌信号盲分离

Blind Separation of Chaotic Signals Based on ICA

现有的混合混沌信号分离方法一般都要利用各个混沌信号的内在性质以及一定的约束。利用混合混沌信号中各源信号的独立性,依据基本ICA估计原理中的极大非高斯性原理,采用基于峭度的不动点分离法对此类混合信号进行分离,实现了此类信号的盲分离。对多种此类混合信号进行分离仿真的结果表明,该方法可以快速有效地分离出混合混沌信号中的各个源信号。

There are some methods that separate the mixing chaotic signals, but they have to use the internal properties of the signals and special constraints. By exploiting the independence of source in the mixing chaotic signals, using the fixed - point ICA based on the kurtosis to separate the mixtures,which is accordance with the ICA estimation principle of maximum nongaussianity. The results by computer simulation indicate that the mixed chaotic signals, by using the method, the source signals can be separated fast and effectively.