基于特征提取的人工智能调制格式识别算法研究现状  

Review on Artificial Intelligence Modulation Format Recognition Algorithm Based on Feature Extraction

作　　者：覃国津 迟楠[1] Qin Guojin;Chi Nan(The Ministry of Education Key Laboratory of Electromagnetic Wave Information Science,Department of Communication Science and Engineering,College of Information Science and Engineering,Fudan University,Shanghai 200433,China)

机构地区：[1]复旦大学信息科学与工程学院通信科学与工程系电磁波信息科学教育部重点实验室,上海200433

出　　处：《激光与光电子学进展》2023年第9期218-227,共10页Laser & Optoelectronics Progress

基　　金：国家自然科学基金杰青项目(61925104);国家自然科学基金重点项目(62031011);鹏城实验室项目(PCL2021A14)。

摘　　要：随着信息领域不断发展,B5G/6G时代接踵而来,数字通信发展迈入更新、更快的阶段。为了适应更高速的通信、更多样的信道,不同的调制格式应运而生,被应用于不同环境。在不同时间、不同信道中改变调制格式,有利于最大化信道利用率。然而,发射端改变的调制格式对于接收机而言是未知的,这不利于不同通信子系统之间的连接,从而影响构建大型的空天地海一体化通信网络,自动调制格式识别算法将成为破局关键。由于人工智能算法在信号处理和基于特征的分类方面具有很大优势,因此基于特征提取的人工智能分类算法在调制格式识别领域拥有巨大的研究价值和实用价值。介绍了几种基于特征提取、人工智能分类的识别方法,并对其在通信领域的应用进行了分析、探讨和总结。With the information field continuously evolving,B5G and 6G have been introduced,and the development of digital communication has entered a stage of faster updates.To adapt to higherspeed communication and more diverse channels,different modulation formats have been proposed and applied in different environments.The purpose of changing modulation format in different time and channels is to maximize the channel utilization.However,the modulation format changed at the transmitter is unknown to the receiver,which does not facilitate the connection between different communication subsystems,thus affecting the construction of a largescale airspacegroundsea integrated communication network.Therefore,modulation format recognition may be key to solving this problem.Because artificial intelligence(AI)has great advantages in signal processing and classification based on feature extraction,modulation format recognition based on feature extraction combined with AI classification algorithm has great research and practical value.This study introduces several methods based on feature extraction and AI and analyzes,discusses,and summarizes their application in the communication field.

关 键 词：调制格式识别 特征提取 人工智能 神经网络 

分 类 号：TN911.3[电子电信—通信与信息系统]