基于人工噪声多用户MIMO系统的窃密算法

人工噪声是一种保障无线通信系统物理层安全的有效策略。现有针对基于人工噪声保密系统的窃密算法仅适用于单用户多输入单输出系统,该文针对这一问题,提出了两种适用于多用户多输入多输出(multi-input multi-output,MIMO)系统的窃密算法。首先通过基于人工噪声的多用户MIMO系统模型与盲源分离线性混合模型的比对,提出了基于盲源分离理论的窃密思路,其次通过改进两种盲源分离算法中的分离矩阵以及选择的非线性函数结构,消除了原始算法估计秘密源信号出现的相位模糊性,同时对每一步迭代中的分离矩阵进行归一化,避免了幅度的不确定性。仿真结果表明,所提两种方法在有效窃密的同时具有较低的计算复杂度,另外,当窃听天线数目越多时,窃密效果越好。

Wiretapping strategies for artificial noise assisted communication in MU-MIMO system

The artificial noise (AN) assisted strategy is invalid to guarantee physical layer security of wireless comunication. The opposite of AN-assisted strategy is investigated only for the sIngle user case. Aiming at this, two wiretapping methods are proposed for AN assisted multiuser multiple-input multiple-output (MU MIMO) system. Firstly, the idea of wiretapping based on blind source separation (BSS) is expounded by the contrast analysis of the AN-assisted MU-MIMO system model and the BSS linear structure. Then, two novel wiretapping algorithms are put forward. The permutation indetermination of secret message is removed through setting limits on the separation matrix and nonlinear function structure, and the phase ambiguity is reduced by the separation matrix normalization. Simulation results show that both of our proposed methods can intercept information effectively with a low computational complexity. In addition, the more antennas eavesdropper equipped the better average BER.