Considerations about the number of coefficients used in Volterra series models for RF Power Amplifiers

Abstract

This article discusses further about the number of coefficients used in the many implementations of Volterra series models, performing a fair comparison between them, using measured data and known figures of merit found in the literature. Among the models implemented, a model that uses sub-sampling techniques allied to Volterra series in a parallel cascade structure aiming accuracy outperforms other variations of Volterra series models, with the drawback of using a higher number of coefficients. Based on the many figures of merit results, this article suggests new trends to be used in RF power amplifier behavioral models.