Weighted criteria for RF power amplifiers identification in wide-band context

Modeling accurately RF power amplifiers (PAs) is an essential first step for predistortion and identification purposes. In WCDMA context, we prove that inaccurate results of identification are achieved because of the inadequacy of the classical criteria of least squares fitting to the WCDMA input pdf. The performances of the identification can be enhanced through a weighted criteria taking into account AM/AM dispersion for low input amplitudes. In this paper, the weighted criteria is validated for 1- carrier WCDMA input and two RF PAs, considering polynomial memoryless model and multibranch polynomial model.