Deep reinforcement learning for near-fieldwideband beamforming in STAR-RIS networks

作     者：Ji WANG Jiayi SUN Wei FANG Zhao CHEN Yue LIU Yuanwei LIU Ji WANG;Jiayi SUN;Wei FANG;Zhao CHEN;Yue LIU;Yuanwei LIU

作者机构：Department of Electronics and Information EngineeringCollege of Physical Science and TechnologyCentral China Normal UniversityWuhan 430079China Beijing National Research Center for Information Science and TechnologyTsinghua UniversityBeijing 100084China Faculty of Applied SciencesMacao Polytechnic UniversityMacao SARChina School of Electronic Engineering and Computer ScienceQueen Mary University of LondonLondon E14NSUK Department of Electrical and Electronic EngineeringThe University of Hong KongHong Kong 999077China 

出 版 物：《Frontiers of Information Technology & Electronic Engineering》 (信息与电子工程前沿(英文版))

年 卷 期：2024年第25卷第12期

页      面：1651-1663页

核心收录：

学科分类：1305[艺术学-设计学（可授艺术学、工学学位）] 13[艺术学] 081104[工学-模式识别与智能系统] 08[工学] 0804[工学-仪器科学与技术] 081101[工学-控制理论与控制工程] 0811[工学-控制科学与工程] 

基　　金：Project supported by the National Natural Science Foundation of China(Nos.62101205 and 62101308) the Key Research and Development Program of Hubei Province,China(No.2023BAB061) 

主　　题：Deep reinforcement learning Near-field beamforming Simultaneously transmitting and reflecting reconfigurable intelligent surface(STAR-RIS) Wideband beam split 

摘      要：A simultaneously transmitting and reflecting reconfigurable intelligent surface(STAR-RIS)assisted multiuser near-field wideband communication system is investigated,in which a robust deep reinforcement learning(DRL)based algorithm is proposed to enhance the users’achievable rate by jointly optimizing the active beamforming at the base station(BS)and passive beamforming at the *** mitigate the beam split issue,the delay-phase hybrid precoding structure is introduced to facilitate wideband *** the coupled nature of the STARRIS phase-shift model,the passive beamforming design is formulated as a problem of hybrid continuous and discrete phase-shift control,and the proposed algorithm controls the high-dimensional continuous action through hybrid action ***,to address the issue of biased estimation encountered by existing DRL algorithms,a softmax operator is introduced into the algorithm to mitigate this *** results illustrate that the proposed algorithm outperforms existing algorithms and overcomes the issues of overestimation and underestimation.