Machine Learning and Intelligent Communications. Third International Conference, MLICOM 2018, Hangzhou, China, July 6-8, 2018, Proceedings

Research Article

Joint D2D Cooperative Relaying and Friendly Jamming Selection for Physical Layer Security

Download
39 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-00557-3_12,
        author={Yijie Luo and Yang Yang and Yanlei Duan and Zhengju Yang},
        title={Joint D2D Cooperative Relaying and Friendly Jamming Selection for Physical Layer Security},
        proceedings={Machine Learning and Intelligent Communications. Third International Conference, MLICOM 2018, Hangzhou, China, July 6-8, 2018, Proceedings},
        proceedings_a={MLICOM},
        year={2018},
        month={10},
        keywords={D2D communications Cooperative relaying Friendly jamming Active eavesdropper Stackelberg game Q-learning},
        doi={10.1007/978-3-030-00557-3_12}
    }
    
  • Yijie Luo
    Yang Yang
    Yanlei Duan
    Zhengju Yang
    Year: 2018
    Joint D2D Cooperative Relaying and Friendly Jamming Selection for Physical Layer Security
    MLICOM
    Springer
    DOI: 10.1007/978-3-030-00557-3_12
Yijie Luo1,*, Yang Yang1,*, Yanlei Duan2,*, Zhengju Yang2,*
  • 1: Army Engineering University of PLA
  • 2: Troop of PLA
*Contact email: yijieluo@sina.com, sheep_1009@163.com, duanyanlei2008@163.com, yangzhengju1001@126.com

Abstract

D2D communications are emerging technologies to improve spectrum efficiency, energy efficiency as well as physical layer security of cellular networks. In most research, D2D users, considered as friendly jammers, can improve the information security of cellular networks. D2D users can also work as cooperative relays between the eNB and the cellular user (CU) to increase the transmission rate and improve the security capacity simultaneously. Considering there exists an active eavesdropper in the cellular network, which can attack as a passive eavesdropper or an active jammer, joint D2D cooperative relaying and friendly jamming selection can enhance the secrecy achievable rate or the transmission rate of CU. We formulate a Stackelberg game between different intelligent agents, and derive the mixed-strategy equilibrium (MSE) via a hierarchical learning algorithm based on Q-learning. Simulation results show that the strategic selections of D2D users and the active eavesdropper are convergent, and the proposed algorithm has a better performance than the random selection method.