浅析基于AI的信道信息预测在6G中的应用  

Application of AI-Based Channel Information Prediction in 6G

作　　者：肖华华 鲁照华[2,3] 胡留军[2,3] XIAO Huahua;LU Zhaohua;HU Liujun(Beijing Jiaotong University,Beijing 100044,China;ZTE Corporation,Shenzhen 518057,China;State Key Laboratory of Mobile Network and Mobile Multimedia Technology,Shenzhen 518055,China)

机构地区：[1]北京交通大学,中国北京100044 [2]中兴通讯股份有限公司,中国深圳518057 [3]移动网络和移动多媒体技术全国重点实验室,中国深圳518055

出　　处：《中兴通讯技术》2024年第3期3-9,共7页ZTE Technology Journal

摘　　要：深入探讨了基于AI的信道信息预测在6G通信系统中的应用,并从信道信息预测的信道模型、信道信息预测方案、数据收集方法、模型监控技术、仿真验证结果,以及基于AI的信道信息预测所面临的机会与挑战这6个维度展开分析。在信道信息预测方案的分析中,广泛探索了多种信道信息预测策略;在数据收集方面,详细分析了不同参考信号传输类型及数据收集范围的影响;在模型监控方面,讨论了3种可能的监控方式。此外,通过仿真验证,充分展示了基于AI的信道信息预测技术的有效性和优越性。The application of AI-based channel information prediction in the 6G communication system is deeply discussed from the following six dimensions:the channel model of channel information prediction,the channel information prediction scheme,the data collection method,the model monitoring technology,validation results,and opportunities and challenges of AI-based channel information prediction.In the analysis of channel information prediction schemes,various channel information prediction strategies are widely explored;in the data collec-tion section,the influence of different reference signal transmission types and data collection ranges is analyzed in detail;in the model moni-toring,three possible monitoring methods are discussed.In addition,the effectiveness and superiority of AI-based channel information pre-diction technology are presented through simulation verification.

关 键 词：多输入多输出 信道信息预测 人工智能 模型监控 模型训练 

分 类 号：TN929.5[电子电信—通信与信息系统] TP18[电子电信—信息与通信工程]