With the rapid development of wireless communications, the industries and academia acknowledge that the millimeter-wave (mm-wave) frequency band is rich in spectral resources. This paper considers the mm-wave train-ground communication system with multiple mobile relays (MRs) in a high-speed rail (HSR) scenario. We use the deep reinforcement learning (DRL) method to solve the beam management problem to maximize the system throughput. First, the inter-beam interference in the MRs scene is modeled due to the effect of the inter-beam angle on the system performance. Second, the maximization problem of system throughput is constructed for the beam management at both the transmitter and receiver. And we proposed the dynamic joint beam management scheme for the base station and the MRs located on the train’s roof based on DRL. Finally, the performance evaluation shows that the proposed scheme displays low computational complexity in the online phase. Additionally, the system throughput performance is close to the ideal optimal beam tracking algorithm, which combines high performance and low complexity, proving the feasibility of using DRL theory for beam management in train-ground communication.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Deep Reinforcement Learning-Based Train-Ground Beamforming Management for Multi-MRs Mm-wave Communication


    Contributors:
    Qiao, Yuanyuan (author) / Niu, Yong (author) / Zhang, Xiangfei (author) / Wang, Ning (author) / Zhong, Zhangdui (author) / Ai, Bo (author)


    Publication date :

    2023-10-10


    Size :

    1298844 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Deep reinforcement learning for hybrid beamforming in multi-user millimeter wave wireless systems

    Lizarraga, Enrique M. / Maggio, Gabriel N. / Dowhuszko, Alexis A. | IEEE | 2021


    Freight train scheduling via decentralised multi-agent deep reinforcement learning

    Bretas, Allan / Mendes, Alexandre / Chalup, Stephan K. | TIBKAT | 2021

    Free access

    Beamforming in Multi-User MISO Cellular Networks with Deep Reinforcement Learning

    Chen, Hongchao / Zheng, Zhe / Liang, Xiaohui et al. | IEEE | 2021


    Train accurate parking control method based on deep reinforcement learning

    ZHANG LEI / ZHANG JIANGUO | European Patent Office | 2022

    Free access

    Train operation scheduling optimization method based on deep reinforcement learning

    LI LIJUAN / YANG XUE / WANG HUAN et al. | European Patent Office | 2023

    Free access