用于低维混沌时间序列预测的一种非线性自适应预测滤波器

在二阶Volterra滤波器基础上，提出了一种用于低维混沌时间自适应预测的非线性自适应预测器。基于最小均方误差准则导出了一种NLMS类型的自适应算法来实时调整这种非线性滤波预测器的系数，仿真实验结果表明：这种线性化的非线性自适应滤波预测器能够有效地预测低维混时间序列，且它的模块化特征更易于VLSI电路实现，具有广泛的工程应用价值。

New nonlinear adaptive predictor for adaptive prediction of low dimensional chaotic time series

A new nonlinear adaptive predictor based on second-order Volterra filters is proposed to make adaptive predictions of chaotic time series. The NLMS-type algorithm, which is derived based on least mean square error, is used to adaptively update this nonlinear predictor's coefficients. Experimental results show that this nonlinear adaptive filter can be used to make adaptive predictions of low dimensional chaotic time series, and the structure of this adaptive predictor and adaptive algorithm are simple and modular, which is convenient for the very large scale integrated( VLSI) circuit implementation.