Publication Type : Conference Proceedings
Publisher : IEEE
Source : IEEE MTT-S Int. Wireless Symp., Nanjing, China
Url : https://ieeexplore.ieee.org/document/9499466
Campus : Bengaluru
School : School of Engineering
Department : Electronics and Communication
Year : 2021
Abstract : This paper reviews the neural network (NN) based digital pre-distortion techniques for linearizing multi-band and multiple-input multiple-output (MIMO) transmitters. The most popular NN-based behavioral models using Deep Neural Networks (DNN), Augmented Neural Networks (ANN), and Shallow Neural Networks (SNN) will be reviewed. The paper will discuss NN-based DPD models using Real-Valued Focused Time-Delay Neural Network (RVFTDNN) and Convolutional Neural Network (CNN) performance and complexity for multi-band and MIMO applications. These models' performance will be assessed in terms of their capability to mitigate the transmitter's distortion and hardware impairments such as crosstalk, PA's nonlinearity, dc offset, and I/Q imbalance for multi-band and MIMO applications.
Cite this Research Publication : P. Jaraut, et al., “Review of the Neural Network based Digital Predistortion Linearization of Multi-Band/MIMO Transmitters,” in IEEE MTT-S Int. Wireless Symp., Nanjing, China