| 168 | 0 | 86 |
| 下载次数 | 被引频次 | 阅读次数 |
在第5代移动通信系统(5G)中，功率放大器(PAs)存在严重的非线性失真和强记忆效应，尤其是对于复杂的高阶调制系统，如64QAM、256QAM等。针对这一问题，提出一种基于MLP模型的智能数字预失真（AI-DPD）系统，用于优化5G NR物理上行共享信道（PUSCH）的链路性能。通过实验对64QAM调制信号下的AI-DPD、无DPD和传统DPD技术进行了性能比较。实验结果表明，AI-DPD技术在所有信噪比（SNR）条件下均展现出最低的误码率（BLER），尤其在BLER为0.01时，AI-DPD技术实现了约1dB的SNR增益。此外，其星座图显示出高度集中和规则的分布，在SNR为30dB时对应的误差向量幅度（EVM）为3.71%，远低于3GPP规定的64 QAM的最小EVM要求8%。相比之下，无DPD技术的星座图分布较为分散，EVM为12.3%，表明信号经功率放大器（PA）后出现严重失真。传统DPD技术的EVM为7.3%。因此，AI-DPD技术在处理PA非线性特性方面相较于传统DPD技术展现了更优的性能。
Abstract：In the 5th generation mobile communication systems(5G), power amplifiers(PAs) exhibit severe nonlinear distortion and strong memory effects, especially for complex modulation systems such as 16QAM, 64QAM, etc. Addressing this issue, this paper proposes an MLP-based digital predistortion(AI-DPD) system to optimize the performance of the 5G NR PUSCH. Through a series of experiments, we compared the performance of AI-DPD, non-DPD, and traditional DPD technologies under 64QAM modulation signals. The experimental results show that AI-DPD technology exhibits the lowest bit error rate(BLER) under all signal-to-noise ratio(SNR) conditions, particularly at BLER of 0.01, where AI-DPD technology achieves approximately 1dB SNR gain. Additionally, its constellation diagram shows a highly concentrated and regular distribution, with an EVM of 3.71%, significantly lower than the minimum VM requirement of 8% for 64QAM set by 3GPP. In contrast, the constellation diagram of non-DPD technology is more dispersed, with an EVM of 12.3%,The EVM of traditional DPD technology is 7.3%, which meets the minimum VM requirement set by 3GPP. Therefore, AI-DPD technology demonstrates superior performance in handling the nonlinear characteristics of PAs compared to traditional DPD technology.
[1]解程杰,马施榆.基于FPGA的并行MMSA数字预失真器设计与实现[J].无线通信技术,2022,31(03):50-53+62.
[2]方俊,叶焱,苏日娜,等.用于5G RF PA线性化的多频段通用数字预失真器[J].微波学报,2022,38(06):90-94.
[3] Raab F H,Asbeck P,Cripps s,et al.Power Amplifiers and Transmitters for RF and Microwave[J].IEEE Transactions on Microwave Theory and Techniques,2002,50(3):814-826.
[4] Pipilos S, Papananos Y, Naskas N, et al. A Transmitter IC for TETRA Systems Based on a Cartesian Feedback Loop Linearization Technique[J].IEEE Journal of Solid-State Circuits,2005,40(3):707-718.
[5] Ibnkahla M, Sombrin J, Castanie F, et al. Neural networks for modeling nonlinear memoryless communication channels[J]. IEEE Trans.Commun,1997,45(7):768-771.
[6] Benvenuto N, Piazza F and Uncini A. A neural network approach to data predistortion with memory in digital radio systems. Proc[J]. IEEE Int. Commun. Conf.Singapore:IEEE.1995:152-156.
[7] M. Rawat et al. Adaptive Digital Predistortion of Wireless Power Amplifiers/Transmitters Using Dynamic Real-Valued Focused TimeDelay Line Neural Networks[J]. IEEE Transactions on Microwave Theory and Techniques,2010, 58(1):95-104.
[8] Yikang Zhang,Yue Li,Falin Liu, et al.Vector Decomposition Based Time-Delay Neural Network Behavioral Model for Digital Predistortion of RF Power Amplifiers.[J].IEEE Access,2019,7(1):91559-91568.
[9] Taijun Liu, Slim Boumaiza,Fadhel M. Ghannouchi.Dynamic Behavioral Modeling of 3G Power Amplifiers Using Real-Valued Time-Delay Neural Networks[J].IEEE Transactions on Microwave Theory and Techniques,2004,52(3):1025-1033.
[10] Farhadi H, Haraldson J, Sundberg M. A Deep Learning Receiver for Non-Linear Transmitter[J]. IEEE Access, 2023,11:2796-2803.
[11] B. Yu et al. Light-Weight AI Enabled Non-Linearity Compensation Leveraging High Order Modulations[J]. IEEE Transactions on Communications,2024, 72(1):539-552.
[12] J. Pihlajasalo et al. HybridDeepRx:Deep Learning Receiver for High-EVM Signals[J]. 2021 IEEE 32nd Annual International Symposium on Personal, Indoor and Mobile Radio Communications(PIMRC), Helsinki,Finland, 2021:622-627.
[13]刘昕,陈文华,吴汇波,等.功放数字预失真线性化技术发展趋势与挑战[J].中国科学:信息科学,2022,52(04):569-595.
[14] 3GPP.TS38.211. Physical Channels and Modulation[S]. France:3GPP Organizational Partners, 2019.
[15] 3GPP. TS38.212. Multiplexing and channel coding[S]. France:3GPP Organizational Partners, 2019.
[16] Further Elaboration on PA Models for NR[S]. document R4-165901, 3GPP, RAN4 Tdoc, 2016.
基本信息:
DOI：
中图分类号:TN722.75;TN929.5
引用信息:
[1]李校林,王嘉航,曾凡琪,等.一种基于MLP的智能数字预失真技术[J].通信与信息技术,2025,No.277(05):1-5.
基金信息: