论文部分内容阅读
混沌信号与确定性小信号叠加生成的混合信号是一更高维的混沌信号 ,因而不能用一般的混沌信号噪声抑制的方法进行分离 .提出了一种这类信号盲分离的方法 .在重构未知的混沌信号的动力方程时 ,充分利用混沌吸引子的几何特性 ,并且限定动力映射为原混沌吸引子所在流形的内部映射 ,从而保证了重构的动力系统方程对应于原混沌信号 ,而不是同样具有混沌特性的混合信号 .然后利用重构的动力方程 ,借用混沌信号中的噪声抑制思想 ,估计出原混沌信号对应的轨道 ,实现信号分离 .通过对Lorenz系统中谐波信号、Henon映象中自回归过程 ,以及脑电信号中谐波信号进行提取的数值实验 ,验证了信号盲分离方法的有效性和可行性 .
The mixed signal generated by the superposition of the chaos signal and the deterministic small signal is a higher-dimensional chaotic signal, and therefore can not be separated by the general method of chaos signal noise suppression. A method of blind separation of such signals is proposed. When the dynamic equation of unknown chaotic signal is used, the geometric characteristics of the chaotic attractor are fully utilized and the dynamical mapping is defined as the internal mapping of the manifold of the original chaotic attractor so as to ensure that the reconstructed dynamic system equation corresponds to the original chaotic signal. Is not a mixed signal with the same chaotic characteristics.According to the idea of noise suppression in the chaotic signal and using the reconstructed dynamic equations, the orbit of the original chaotic signal can be estimated and the signal separation can be achieved.Based on the harmonic signals in the Lorenz system, In the autoregressive process, as well as the extraction of harmonic signals in EEG, a numerical experiment is carried out to verify the effectiveness and feasibility of the signal blind separation method.