ZTE Communications ›› 2025, Vol. 23 ›› Issue (1): 71-77.DOI: 10.12142/ZTECOM.202501009

• Research Papers • Previous Articles     Next Articles

A Basis Function Generation Based Digital Predistortion Concurrent Neural Network Model for RF Power Amplifiers

SHAO Jianfeng1(), HONG Xi1, WANG Wenjie1, LIN Zeyu2, LI Yunhua2   

  1. 1.Xi’an Jiaotong University, Xi’an 710049, China
    2.ZTE Corporation, Shenzhen 518057, China
  • Received:2023-09-23 Online:2025-03-25 Published:2025-03-25
  • About author:SHAO Jianfeng (sjf1996717@stu.xjtu.edu.cn) received his BS degree in electronic information engineering from the School of Information Engineering, North China University of Water Resources and Electric Power in 2018, MS degree in circuits and systems from the School of Physics and Electronic Engineering, Ningxia University, China in 2021. He is pursuing a PhD degree in information and communication engineering at Xi'an Jiaotong University, China. His research interests include array signal processing and digital pre-distortion.
    HONG Xi received his BS degree in information engineering from the School of Electronics and Information Engineering, Xi'an Jiaotong University, China in 2012, and PhD degree in information and communication engineering from the School of Information and Communication Engineering, Xi'an Jiaotong University in 2022. He is currently an engineer at the School of Information and Communication Engineering, Xi'an Jiaotong University. His research interests include array signal processing, signal processing in communication systems, multipath mitigation in navigation, and GNSS physical layer interference detection.
    WANG Wenjie received his BS, MS, and PhD degrees in information and communication engineering from Xi'an Jiaotong University, China in 1993, 1998, and 2001, respectively, where he is currently a professor. His main research interests include information theory, broadband wireless communications, signal processing in communication systems, and array signal processing.
    LIN Zeyu received his BS and MS degrees at the School of Information and Communication Engineering, Xi'an Jiaotong University, China in 2018 and 2021, respectively. He is currently a senior RF algorithm architect in the RHP department of ZTE Corporation. His research interests include PA nonlinear system behavioral modeling and RF systems linearization.
    LI Yunhua received his BE degree in communications engineering from Shandong University (Weihai), China in 2012 and PhD degree in military communications from the School of Telecommunications Engineering, Xidian University, China in 2017. He is currently a senior engineer of RF algorithms in ZTE Corporation. His research interests include communication signal processing, wireless communications, digital predistortion modeling of nonlinear systems, and more.
  • Supported by:
    ZTE Industry?University?Institute Cooperation Funds(HC?CN?20220722010)

Abstract:

This paper proposes a concurrent neural network model to mitigate non-linear distortion in power amplifiers using a basis function generation approach. The model is designed using polynomial expansion and comprises a feedforward neural network (FNN) and a convolutional neural network (CNN). The proposed model takes the basic elements that form the bases as input, defined by the generalized memory polynomial (GMP) and dynamic deviation reduction (DDR) models. The FNN generates the basis function and its output represents the basis values, while the CNN generates weights for the corresponding bases. Through the concurrent training of FNN and CNN, the hidden layer coefficients are updated, and the complex multiplication of their outputs yields the trained in-phase/quadrature (I/Q) signals. The proposed model was trained and tested using 300 MHz and 400 MHz broadband data in an orthogonal frequency division multiplexing (OFDM) communication system. The results show that the model achieves an adjacent channel power ratio (ACPR) of less than –48 dB within a 100 MHz integral bandwidth for both the training and test datasets.

Key words: basis function generation, digital predistortion, generalized memory polynomial, dynamic deviation reduction, neural network