基于實(shí)測信道的AI賦能無線通信：信道反饋

作者：Jiajia Guo1, Xiangyi Li1, Muhan Chen1, Peiwen Jiang1, Tingting Yang2, Weiming Duan2,Haowen Wang3, Shi Jin1, Quan Yu2

單位：1. National Mobile Communications Research Laboratory, Southeast University; 2. Peng Cheng Laboratory; 3. Laboratory of Broadband Wireless Technology, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences

近年來，人工智能（Artificial Intelligence，AI）在信號處理、信道估計、編碼設(shè)計等通信領(lǐng)域取得了重大突破，打破了傳統(tǒng)通信系統(tǒng)的設(shè)計瓶頸，作為一項突破性技術(shù)使得智能通信成為未來通信系統(tǒng)研究的熱門方向之一。基于深度學(xué)習(xí)（Deep Learning，DL）的信道狀態(tài)信息（Channel State Information，CSI）反饋技術(shù)因其突出的性能優(yōu)勢得到了廣泛關(guān)注，但目前相關(guān)研究僅使用模擬生成的數(shù)據(jù)集來訓(xùn)練和測試，無法保證AI算法在實(shí)際通信系統(tǒng)中仍然具有良好的性能。

為了探索AI在通信系統(tǒng)中的實(shí)際性能，鵬城實(shí)驗(yàn)室等單位組織了全國人工智能大賽（NAIC）“AI+無線通信”賽道，初賽賽題為“基于AI的無線通信信道的壓縮及恢復(fù)”。本文詳細(xì)描述了該比賽的信道數(shù)據(jù)采集過程，同時為該比賽提供了一個基于DL的CSI反饋參考架構(gòu)：QuanCsiNet，實(shí)現(xiàn)了真實(shí)信道場景采集的高維信道數(shù)據(jù)的壓縮、量化、反饋和重建，為AI在未來通信系統(tǒng)中的實(shí)際部署和使用奠定了基礎(chǔ)。

真實(shí)信道場景為圖1所示的辦公室場景，發(fā)射機(jī)在圖中所示位置固定，接收機(jī)沿紅點(diǎn)軌跡運(yùn)動。對測量得到的真實(shí)信道數(shù)據(jù)進(jìn)行圖2所示的預(yù)處理，得到最終使用的數(shù)據(jù)集。

圖1 實(shí)測信道數(shù)據(jù)場景示意圖

圖2 實(shí)測信道數(shù)據(jù)處理流程

QuanCsiNet的網(wǎng)絡(luò)結(jié)構(gòu)如下圖所示。編碼器用于對CSI進(jìn)行特征提取和壓縮，量化模塊用于將壓縮測量值用有限位表示，轉(zhuǎn)化為比特流便于實(shí)際系統(tǒng)存儲和傳輸；逆量化模塊用于將比特流恢復(fù)成壓縮測量值，譯碼器用于特征解壓縮和信道恢復(fù)。

圖3 QuanCsiNet網(wǎng)絡(luò)結(jié)構(gòu)

具體來說，文章

• 首次針對實(shí)測信道數(shù)據(jù)進(jìn)行處理，提出了一種真實(shí)信道場景下的CSI反饋架構(gòu)，為后續(xù)研究提供了可擴(kuò)展的參考設(shè)計。

• 引入了量化和逆量化模塊，將反饋測量值轉(zhuǎn)化為比特流，符合實(shí)際系統(tǒng)存儲傳輸要求。

• 評估了基于DL的CSI反饋方案在實(shí)際信道環(huán)境中的性能，推動后續(xù)研究和實(shí)際部署。

整體而言，本文對真實(shí)信道場景下采集的信道數(shù)據(jù)進(jìn)行處理，設(shè)計了一種以比特流形式進(jìn)行反饋的基于DL的CSI反饋架構(gòu)，并衡量了其重建性能與復(fù)雜度，以期更多研究者為智能通信的實(shí)際應(yīng)用作出貢獻(xiàn)。

論文下載鏈接：（請戳此處）

引用格式：Jiajia Guo, Xiangyi Li, Muhan Chen, Peiwen Jiang, Tingting Yang, Weiming Duan, Haowen Wang, Shi Jin, Quan Yu, “AI Enabled Wireless Communications with Real Channel Measurements: Channel Feedback”, Journal of Communications and Information Networks, vol. 5, no. 3, pp. 310-317, Sep. 2020.

本文由論文作者供稿。

歡迎課題組投遞成果宣傳稿（可在文章下方留言聯(lián)系我們）！

作者簡介

Jiajia Guo（郭佳佳）received the B.S. degree from Nanjing University of Science and Technology, Nanjing, China, in 2016, and the M.S. degree from University of Science and Technology of China, Hefei, China, in 2019. He is currently working towards the Ph.D. degree in information and communications engineering, Southeast University, China. His research interests currently include, deep learning, neural network compression, massive MIMO, and machine learning in communications.

Xiangyi Li（李湘宜）received the B.S. degree from School of Mathematics, Tianjin University, Tianjin, China, in 2017, and the M.S. degree from Centre for Applied Mathematics, Tianjin University, in 2020. She is currently working toward the Ph.D. degree in information and communications engineering, Southeast University, China. Her main research focuses on deep learning application in wireless communication and massive MIMO systems.

Muhan Chen（陳慕涵）received the B.S. degree from the School of Information Science and Engineering, Southeast University, Nanjing, China, in 2019. She is currently working toward the M.S. degree with the School of Information Science and Engineering, Southeast University, Nanjing, China. Her research interests center around deep learning applications in wireless communication systems.

Peiwen Jiang（姜培文）received the B.S. degree from Southeast University, Nanjing, China in 2019. He is currently working toward the Ph.D. degree with the School of Information Science and Engineering, Southeast University. His research interests include deep learning-based channel estimation and signal detection in communications.

Tingting Yang（楊婷婷）received her B.Sc. and Ph.D. degrees from Dalian Maritime University, China, in 2004 and 2010, respectively. She is currently a Research Professor at Peng Cheng Laboratory, China. Her research interests are in the areas of maritime wideband communication networks, AI-empowered wireless communications. She serves as the Associate Editor-in-Chief of the IET Communications, as well as the Advisory Editor for SpringerPlus.

Weiming Duan（段為明）is now a Senior Engineer in Peng Cheng Laboratory. He received his M.S. degree in communication and information system from the University of Electronic Science and Technology of China (UESTC) in 1999. In the same year, he joined Huawei Wireless Research Department in Shanghai and has worked there for 20 years. He has worked on baseband algorithm for 3G/4G, advanced receiver for 4G, waveform concept research for 5G, and has also been deeply involved in low-level algorithm library optimization to speed up largescale system simulation.

Haowen Wang（王浩文）is a Senior Engineer of Laboratory of Broadband Wireless Technology, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences (SIMIT). He received his B.S. and M.S. degrees from EE department and College of Software of Fudan University. In SIMIT, Haowen is a leader of wireless technology R&D group. He has many years of experience in the test and verification for the new technologies of wireless communications. His job and research interests include RF data acquisition, channel measurement, verification and test solution.

Shi Jin（金石）[corresponding author] received his B.S. degree in communications engineering from Guilin University of Electronic Technology, Guilin, China, in 1996, his M.S. degree from Nanjing University of Posts and Telecommunications, Nanjing, China, in 2003, and his Ph.D. degree in information and communications engineering from Southeast University, Nanjing, in 2007. From June 2007 to October 2009, he was a Research Fellow with the Adastral Park Research Campus, University College London, London, U.K. He is currently with the Faculty of the National Mobile Communications Research Laboratory, Southeast University. His research interests include space time wireless communications, random matrix theory, and information theory. He serves as an Associate Editor for the IEEE Transactions on Wireless Communications, IEEE Communications Letters, and IET Communications. He and his coauthors have been awarded the 2011 IEEE Communications Society Stephen O. Rice Prize Paper Award in the field of communication theory and the 2010 Young Author Best Paper Award by IEEE Signal Processing Society.

Quan Yu（于全）received his B.S. degree in radio physics from Nanjing University, China, in 1986, his M.S. degree in radio wave propagation from Xidian University, China, in 1988, and his Ph.D. degree in fiber optics from the University of Limoges, France, in 1992. He is currently a Research Professor at Peng Cheng Laboratory. His main areas of research interest are the architecture of wireless networks and cognitive radio. He is an Academician of the Chinese Academy of Engineering (CAE) and the founding Editor-in-Chief of the Journal of Communications and Information Networks (JCIN).

雷鋒網(wǎng)雷鋒網(wǎng)