An Improved Rational Model with Memory for the Digital Baseband Predistortion of Power Amplifier

Abstract

The use of models described by the ratio of two polynomial approximations with memory has been receiving particular attention by the wireless communication community. Rational models with memory require a lower number of coefficients and work better with extrapolation data than polynomial approximations. This work proposes a novel rational model with memory that, in comparison with previous approaches having the same number of parameters, can increase the modeling accuracy. The novelties of the proposed rational model, in comparison with the best available rational model, are the use of a larger set of truncation factors and the inclusion of additional terms.