Predicting low-dimensional chaotic time series using Volterra adaptive filers

Volterra adaptive filter is used to predict low-dimensional chaotic time se ries based on the state space reconstruction of delay-coordinate embedding of dynamic system. It is shown, through experiments of predicting eight kin ds of low-dimensional chaotic series using second-order Volterra adaptive f ilters, that Volterra adaptive filters can accurately predict these chaotic series when the length N-l of the Volterra filler is long enough, and the choice of N-l is related to D-2 and smoothness of chaotic map.