Behavioral Modeling of RF Power Amplifiers Using Adaptive Recursive Polynomial Functions

In this paper a novel adaptive approach is used to model the behavior of a nonlinear RF power amplifier with memory. The recursive input to the model enables a reduction of the number of coefficients required to model the system. Furthermore the equation-error approach used does not suffer from the convergence problems experienced using other recursive adaptive techniques. Validation of the approach was achieved using separate training and validation signals from various nonlinear power amplifiers. The improved performance of the proposed approach was measured using goodness of fit statistics and compared against existing methods