Systematic pruning of Volterra series using Wiener G-functionals for power amplifier and predistorter modeling

The Volterra series is a well-established technique for modeling nonlinear dynamic system. Their wide adoption in predicting the behavior of Radio Frequency Power Amplifiers has been significantly hindered by the large number of coefficients involved especially when they exhibit high order of nonlinearity and strong memory effects. In this paper, we present a new pruning approach based on Wiener G-functionals which is sought for the selection of reduced number of dominant kernels in the Volterra series.