中国物理B ›› 2016, Vol. 25 ›› Issue (1): 10202-010202.doi: 10.1088/1674-1056/25/1/010202
Wen-Bo Wang(王文波), Xiao-Dong Zhang(张晓东), Yuchan Chang(常毓禅), Xiang-Li Wang(汪祥莉), Zhao Wang(王钊), Xi Chen(陈希), Lei Zheng(郑雷)
Wen-Bo Wang(王文波)1,2, Xiao-Dong Zhang(张晓东)2, Yuchan Chang(常毓禅)3, Xiang-Li Wang(汪祥莉)4, Zhao Wang(王钊)5, Xi Chen(陈希)5, Lei Zheng(郑雷)6
摘要: In this paper, a new method to reduce noises within chaotic signals based on ICA (independent component analysis) and EMD (empirical mode decomposition) is proposed. The basic idea is decomposing chaotic signals and constructing multidimensional input vectors, firstly, on the base of EMD and its translation invariance. Secondly, it makes the independent component analysis on the input vectors, which means that a self adapting denoising is carried out for the intrinsic mode functions (IMFs) of chaotic signals. Finally, all IMFs compose the new denoised chaotic signal. Experiments on the Lorenz chaotic signal composed of different Gaussian noises and the monthly observed chaotic sequence on sunspots were put into practice. The results proved that the method proposed in this paper is effective in denoising of chaotic signals. Moreover, it can correct the center point in the phase space effectively, which makes it approach the real track of the chaotic attractor.
中图分类号: (Fourier analysis)