Behavioral modeling and linearization

Abstract

This chapter describes the linearization of power amplifiers (PAs) used in radio frequency (RF) transmitters. An overview of wireless communications is provided to establish the context of how the transmitter is used and the importance of the PA. Measures and models of the PA nonlinearity are reviewed, after which the core topic of PA linearization is presented. Selected techniques for PA linearization are discussed. The most popular is digital predistortion (DPD), where the correction signal is applied in the digital domain prior to the PA. Also covered are adaptive methods used to estimate the DPD coefficients that optimize the system. Feedforward compensation is presented as an analog technique for linearization, where the correction is applied in the RF domain after the PA. The chapter is concluded with a section on more detailed discussions of selected topics. Topics include the use of both odd and even polynomial terms in the modeling of PA nonlinearity and predistortion, the effect of PA saturation on the estimation of DPD coefficients, modifications used to make descent-based estimators more robust, batch processing for coefficient estimation, and the recursive least squares (LS) (RLS) estimator.