Cascaded Model of Nonlinear Operator for Digital Predistortion with Memory

Abstract

Nonlinear distortion with memory perturbing the wideband signal in power amplifiers (PAs) is challenging both for analytical treatment and implementation of practical PA linearizers. In the present letter we focus on a simplified cascaded model of the PA memory nonlinearity that enables to approach such problems as the effect of modeling and identification errors and undersampling on the digital predistortion performance. Based on the presented analysis, we propose a cascaded nonlinearity model for predistorter with a 2-step identification algorithm using look-up table to improve stability of the model identification. Proposed model is shown to have significant performance gain compared to conventional memory polynomial nonlinearity model for specific models of PA distortion.