Multi-band behavioral modeling of power amplifier using carrier frequency-dependent time delay neural network model

Abstract

In this paper, a novel carrier frequencydependent time delay neural network (CF-dependent TDNN) model for behavioral modeling of a memory power amplifier (PA) for multi-band operation is proposed. Traditional methods such as memory polynomial and time delay neural network characterize the nonlinearity of a PA in a single wide band with multiple carriers. However, these methods cannot accurately capture the dynamic nonlinearity in multi-band transmitters. The proposed CF-dependent TDNN model can easily capture and accurately characterize the dynamic nonlinearity in a PA even when the PA is driven in more than one band of operation. Further, the CF-dependent TDNN model is developed by simultaneous training using the PA data from multiple bands, thereby making the model easier to develop and capture the nonlinearity and memory effects of a PA in multiple bands. The proposed method is validated and confirmed with simulation of a Cree RF GaN HEMT PA.