Particle swarm optimization assisted B-spline neural network based predistorter design to enable transmit precoding for nonlinear MIMO downlink

作     者：Chen, Sheng Ng, Soon Xin Khalaf, Emad Morfeq, Ali Alotaibi, Naif 

作者机构：Univ Southampton Sch Elect & Comp Sci Southampton SO17 1BJ Hants England King Abdulaziz Univ Elect & Comp Engn Dept Fac Engn Jeddah 21589 Saudi Arabia 

出 版 物：《NEUROCOMPUTING》 (神经计算)

年 卷 期：2021年第458卷

页      面：336-348页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Deanship of Scientific Research (DSR) at King Abdulaziz University, Jeddah [RG-3-135-40] DSR 

主　　题：Particle swarm optimization B-spline neural network Nonlinear inversion Nonlinear high power amplifier Predistorter Multi-input multi-output Downlink precoding 

摘      要：For the multiple-input multiple-output (MIMO) downlink employing high-order quadrature amplitude modulation signaling and with nonlinear high power amplifiers (HPAs) at base station transmitter, the existing precoding designs relying on the linear MIMO channel can no longer work. We propose an effi-cient and accurate predistorter design to enable transmit precoding for nonlinear MIMO downlink. Specifically, we obtain the closed-form least squares estimates of the nonlinear HPA s amplitude and phase response using two B-spline neural networks during training. The estimated HPA s phase response automatically yields the estimate of the predistorter s phase response. Based on the B-spline neural net-work estimate of the HPA s amplitude response, we construct a B-spline neural network model for the predistorter amplitude response, and we adopt a particle swarm optimization (PSO) algorithm to solve this highly nonlinear optimization problem. Using our accurate predistorter estimate to pre-compensate for the nonlinear distortions of the transmit HPAs, a standard full-digital transmit precoding design can readily be adopted to combat the MIMO channel interference. A simulation study is conducted to demonstrate the effectiveness of our proposed PSO assisted predistorter design. (c) 2021 Elsevier B.V. All rights reserved.