Envelope and Frequency-Added Neural Network Digital Predistortion of Envelope Tracking PAs for Flexible Spectrum Scenarios

The intelligent communication system dynamically adjusts the configuration of transmitter according to the current environment and user needs for efficiency. In this paper, an envelope and frequency-added neural network digital predistortion (EFANN-DPD) method is proposed to enhance the efficiency of the envelope tracking power amplifier (ET PA) and improve its linearity. The system automatically captures the influence of frequency variations on the PA characteristics and integrates them into the neural network. The proposed model can linearize the PA in a time-varying environment without the need for coefficient updating. In order to adapt to the physical characteristics of ET PAs, the envelope input is also considered and integrated into the neural network. Experimental results show that EFANN-DPD maintains good linearity when the operating frequency of the PA changed in real time.