基于Volterra级数的RLS自适应算法的混沌时间序列预测

RLS Self-adaptive Prediction for Chaotic Time Series Based on Volterra Series

根据混沌序列产生的确定性和非线性机制,基于Volterra级数和RLS算法,提出了一种少参数二阶非线性滤波算法用于混沌时间序列的自适应预测。仿真结果表明,这种非线性自适应滤波器能有效地预测一些超混沌序列,而且该滤波器的一步均方误差性能明显高于其他基于Volterra级数的NLMS算法,表明该算法具有良好的收敛性能。

According to the deterministic and nonlinear mechanism of the generation of chaotic sequences,an oligo-parameter second-order nonlinear filter algorithm based on Volterra series and the RLS algorithm is presented for adaptive prediction of chaotic time sequences.Simulation results show that this nonlinear adaptive filter can effectively predict some micro-chaotic sequence.Moreover,one-step mean-square error characteristic of the filter is apparently higher than the other NLMS algorithms based on Volterra series,which shows that the algorithm has good convergence performance.