Beamforming Design for Integrated Communication and Jamming Systems with Unknown CSI:A Hybrid Data-Driven and Model-Based Approach  

作　　者：Liu Jiteng Ding Guoru Xu Yitao Wang Haichao Gu Jiangchun 

机构地区：[1]College of Communications Engineering,Army Engineering University of PLA,Nanjing 210007,China

出　　处：《China Communications》2025年第4期81-99,共19页中国通信(英文版)

基　　金：supported by the National Natural Science Foundation of China(No.62171462,No.62401626,No.62271501);the Key Technologies R&D Program of Jiangsu(Prospective and Key Technologies for Industry)under Grants BE2023022 and BE2023022-4;the Natural Science Foundation of Jiangsu Province(No.BK20240200)。

摘　　要：The integrated communication and jamming(ICAJ)system recently has been proposed to enable communication and jamming(C&J)to reinforce each other in one system.By exploiting the diversity gain of multiple input multiple output(MIMO)technology,a specific implementation form of ICAJ system,called communication-aided collaborative jamming system,is designed to transmit C&J signals at the same time and frequency.Different from previous studies which overlook the jamming prior information acquisition process and assume that the prior information is perfect or with bounded error,this paper takes the non-cooperative characteristics of jamming and the consequent difficulty in prior information acquisition into consideration.To analyze the tradeoff between C&J,the integration metric is proposed and then the corresponding system design problem is formulated.However,the non-convexity of problem and the lack of jamming prior information make the optimization tricky.In this case,blind channel estimation(BCE)is introduced to obtain an approximate channel state information(CSI)without interacting with jamming targets and then the neural network embedded with system performance calculation model is developed to establish the correspondence between the estimated CSI and optimal beamforming design.Furthermore,a hybrid data-driven and model-based approach,blind channel estimation-deep learning(BCEDL),is proposed to accomplish the beamforming design based on unsupervised learning for ICAJ system in non-cooperative scenarios.The simulation results show that the BCE-DL algorithm outperforms the conventional algorithms in the presence of CSI estimation errors and is a flexible approach which takes the best of both data-driven and model-based methods to design the ICAJ system.

关 键 词：blind channel estimation deep learning integrated communication and jamming(ICAJ) multiple input multiple output(MIMO) 

分 类 号：TN92[电子电信—通信与信息系统] TN911.7[电子电信—信息与通信工程]