一种基于三阶Volterra滤波器的混沌时间序列自适应预测方法

基于Takens的相空间延迟坐标重构,研究了用于混沌信号预测的三阶Volterra滤波器的一种乘积耦合近似实现结构,并应用于典型的低维混沌时间序列和具有高维混沌特性的EEG信号的预测.数值研究表明:这种滤波器结构对于低维混沌时间序列的预测精度可以比二阶Volterra滤波器提高103倍,而且能够较好地对一些具有高维混沌特性的EEG信号进行预测 关键词： 混沌 非线性自适应预测 三阶Volterra滤波器 electroencephalography信号

A method based on the third-order Volterra filter for adaptive predictions of chaotic time series

Based on the Takens' delay coordinate phase reconstruct, we study the third order Volterra filter which is used to make adaptive predictions of chaotic signals. It is approximately implemented by a product coupling configuration; and this filter is used to predict typical low dimensional chaotic time series and high dimensional chaotic electro encephalography(EEG) signal. Simulation results show that: this filter has aprecision 10 3 times higher than the second order Volterra filter when it is used to make predictions of low dimensional chaotic time series. It can be successfully used to make predictions of some high dimensional chaotic EEG signal.