摘要
确定的非线性系统能够产生貌似噪声的混沌,本文分析了混沌和白噪声的统计特性及时频特性,指出利用传统的时频仪器是无法区分噪声和混沌的.并且提出一种可区分混沌与噪声的"相关空间法",基于Logist方程的迭代混沌序列是由确定性方程产生的,具有较小的自由度,必然存在短期可预测性;而白噪声却是由大量的偶然因素产生的,具有较多的自由度,不存在可预测性.进而提出由相互关联的临近序列组成相关空间,将迭代序列投影于该相关空间,从而区分混沌和噪声.
Nonlinear deterministic system may result in chaos which appears to be noise. In this paper, statistical characteristic and time-frequency characteristic of chaos and white noise are analyzed and the outcome is that traditional instruments can not be used in distinguishing chaos and noise. The method to distinguish chaos and noise is presented. Legist iterated chaotic sequence is resulted from deterministic equation. It has small degree of freedom and possesses short time predictability, while white noise has no such properties. So correlation space which composed of the near associated sequence is suggested, and chaotic time series and white noise can be distinguished from the projection in the correlation space.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2002年第4期419-423,共5页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金(50077016
60176020)
博士点资金(98069828)
