D2D使能蜂窝网络中存在主动窃听者的物理层安全研究

D2D通信作为5G移动通信的关键技术，以其能有效提升系统能量效率和频谱效率而备受瞩目。在系统中存在被动窃听者的情况下，D2D带来的干扰也被视为友好干扰来提升蜂窝系统的物理层安全性能。但如果存在的是主动窃听者， D2D带来的干扰就会影响蜂窝系统安全稳健的传输性能。因此本文考虑在系统中存在可以智能选择被动窃听或主动干扰的攻击者的情况下，通过选择最佳的D2D用户并采取协同中继或友好干扰的策略来提升在被动窃听情况下蜂窝网络的安全可达速率和在主动干扰下的数据传输速率，来改进蜂窝系统通信的鲁棒性和安全性。本文将合法用户和主动窃听者之间的相互作用建模为非零和的非协作混合博弈，分析了此混合博弈的纳什均衡的存在性，并通过基于fictitious play(虚拟对策)的算法获得了其纳什均衡解。仿真结果表明该算法优于传统的nearest neighbor(最近距离)算法，并且联合考虑两种D2D协作方式的策略优于只将D2D用户视为友好干扰或是协同中继的策略。 

Research on physical layer security in D2D enabled cellular networks with an active eavesdropper

D2D communication, as a key technology of the 5G mobile communication system, draws a lot of attentions for its ability to enhance energy efficiency and spectrum efficiency of cellular systems. Even the interference that it brings can be regarded as “friendly” when there is a passive eavesdropper in systems, since it can improve the physical layer security of cellular users. But if there is an active eavesdropper in systems, it will affect the secure and robust communication performance. So in this paper, in order to improve robustness and security of D2D enabled cellular networks, considering there exists an attacker who has two attack modes: passively eavesdropping or actively jamming, the secrecy achievable rate and transmission rate of legitimate user under different attacked scenario are enhanced by the assistance of the selected D2D pair using joint friendly jamming and cooperative relaying method. The interaction between the legitimate user and the attacker is formulated as a non-zero non-cooperative game, and the existence of the mixture strategy is analyzed. The MSNE (mixture strategy Nash equilibrium) is derived via the algorithm based on fictitious play. Simulation results show that the proposed algorithm outperforms the classical nearest neighbor algorithm, and the strategy joint considering two collaborative ways of the selected D2D pair works better than the strategy that only considering it to be a cooperative relay or a friendly jammer.