AI-Based transceivers for massive MIMO wireless communications

用于大规模 MIMO 无线通信的基于 AI 的收发器

• 批准号: 555846-2020
• 负责人: LeNgoc, Tho
• 金额: $ 2.68万
• 依托单位: McGill University
• 依托单位国家: 加拿大
• 项目类别: Alliance Grants
• 财政年份: 2020
• 资助国家: 加拿大
• 起止时间: 2020-01-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=704944
• 关键词: AI; Based; transceivers; massive; MIMO

Demands for throughput and latency improvements in 5G-and-beyond wireless networks continue to grow at an increasing pace, boosted by new multi-media and social network applications, and new types of massive connectivity such as machine-to-machine (M2M). To meet these growing demands, the use of massive multiple-input multiple-output (M-MIMO) systems operating in the mmWave and THz bands has been recently considered. However, the design of multi-user (MU) M-MIMO transceivers with large-scale antenna arrays (LSAAs) for optimal operation under complex radio conditions and user patterns poses formidable problems, seemingly intractable by traditional means. Artificial intelligence (AI) and machine learning (ML) have the potential to overcome this challenge by exploiting all possible side information relevant to intelligent transceiver signal processing and operations. This proposed 3-year research in collaboration with InterDigital Canada is aimed at studying and developing new AI-enabled transceiver design techniques and algorithms for MU-M-MIMO communications in sub-6GHz, mmWave and THz bands. Cutting-edge ML approaches, especially deep learning (DL) architectures that employ multiple hidden processing layers to learn complex non-linear relationships from large amount of data, will be a corner stone of our work. In recent years, InterDigital has taken a leading role in the development of new technologies for 5G wireless and, as part of this initiative, it will support our research activities through cash and in-kind contributions. AI-enabled MU-M-MIMO transceivers in sub-6GHz, mmWave and THz bands will bridge the performance gap between existing 4G technology and future 5G-and-beyond wireless networks. The proposed research is of great importance to InterDigital and Canada: it will empower the company with critical knowledge and expertise in the key area of AI-enabled wireless networking, and help our country maintain a strategic position in the development, standardization and commercialization of advanced information and communications technologies.

在新的多媒体和社交网络应用以及机器对机器 (M2M) 等新型大规模连接的推动下，5G 及更高版本无线网络对吞吐量和延迟改进的需求继续以越来越快的速度增长。为了满足这些不断增长的需求，最近考虑使用在毫米波和太赫兹频段运行的大规模多输入多输出（M-MIMO）系统。然而，为了在复杂的无线电条件和用户模式下实现最佳操作而设计具有大规模天线阵列 (LSAA) 的多用户 (MU) M-MIMO 收发器会带来棘手的问题，而传统方法似乎难以解决。人工智能 (AI) 和机器学习 (ML) 有潜力通过利用与智能收发器信号处理和操作相关的所有可能的辅助信息来克服这一挑战。这项与 InterDigital Canada 合作的为期 3 年的研究旨在研究和开发新的人工智能收发器设计技术和算法，用于 6GHz 以下、毫米波和太赫兹频段的 MU-M-MIMO 通信。尖端的机器学习方法，特别是采用多个隐藏处理层从大量数据中学习复杂非线性关系的深度学习 (DL) 架构，将成为我们工作的基石。近年来，InterDigital 在 5G 无线新技术的开发中发挥了主导作用，作为该计划的一部分，它将通过现金和实物捐助支持我们的研究活动。支持 AI 的 6GHz 以下、毫米波和太赫兹频段的 MU-M-MIMO 收发器将缩小现有 4G 技术与未来 5G 及以上无线网络之间的性能差距。 拟议的研究对于InterDigital和加拿大都非常重要：它将赋予该公司在人工智能无线网络关键领域的关键知识和专业知识，并帮助我国在先进技术的开发、标准化和商业化方面保持战略地位。信息和通信技术。