大规模MIMO系统中的物理层安全机制研究

发布时间：2018-01-06 01:30

本文关键词：大规模MIMO系统中的物理层安全机制研究　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着移动互联网与物联网的迅猛发展,各类应用对移动网络的容量、速率、能量效率的要求不断的提高。在第五代移动网络(5G)的研究过程中,大规模MIMO技术被认为是5G网络最重要的支撑技术之一。本论文针对大规模MIMO技术中的物理层安全隐患展开研究,提出了基于随机符号方法的和基于子空间方法的两种有效的主动窃听行为检测机制。基于随机符号方法的主动窃听行为检测机制,以专用随机符号的统计特性为基础,构建合理的检验统计量,通过分析不存在主动窃听行为时检验统计量的统计分布规律设计安全区域,从而进行是否存在主动窃听行为的假设检验。将多次独立实验结果融合来提高整体检测性能。同时,增加功率辅助检测机制以避免更加智能的主动窃听用户的逃避检测机制的行为。算法理论检测性能分析和仿真实验均证明基于随机符号方法的检测机制能有效检测主动窃听行为,其检测性能随基站天线数、基站接收信号信噪比、主动窃听信号功率、随机符号序列长度的增长而显著提升,且优于现有同类检测算法。最后,检测算法能在一定程度上提升系统的安全容量。基于子空间方法的主动窃听行为检测机制,以信号子空间方法为基础,随着基站接收导频信号矩阵的维度和样本量的增大,其经验谱分布会逐渐收敛于其极限谱分布,因此可将矩阵的极限谱分布作为理论依据来判断其经验谱分布是否正常,即是否包含非法的主动窃听信号分量。算法首先对合法信号分量使用大维随机矩阵理论求解其极限谱分布,然后计算检验统计量的理论值。利用大维随机矩阵理论去除特征谱中包含的噪声分量,而后进行假设检验以判断是否存在主动窃听行为。仿真实验证明,基于子空间方法的检测机制能有效检测主动窃听行为,其性能随基站与合法用户天线数的比值、采样点数、基站接收信号信噪比、主动窃听信号功率的增大显著增强,且优于同类对比算法。
[Abstract]:With the rapid development of mobile Internet and the Internet of things, the capacity of the mobile network, all kinds of application rate, energy efficiency requirements increase. In the fifth generation mobile network (5G) in the course of the study, large scale MIMO 5G network technology is considered to be the most important supporting technologies. This thesis focuses on the physical layer of large scale MIMO in safety research, put forward based on random symbol method and detection mechanism based on two effective active eavesdropping subspace method. The detection mechanism of active eavesdropping method based on random symbols, with special statistical characteristics of random symbols based on construction of test statistics reasonably, through the analysis of test statistics distribution of design safety area statistics do not exist to active eavesdropping, hypothesis testing the existence of active eavesdropping. The number of independent experiments To improve the overall detection performance of fusion. At the same time, increase the power assisted detection mechanism to avoid the escape detection mechanism more intelligent active eavesdropping user behavior. The theory of detection algorithm performance analysis and simulation experiments show that the detection mechanism of random symbol method can effectively detect the active eavesdropping behavior based on the detection performance with base station antennas, the base station receives a signal SNR, active eavesdropping signal power, the length of the random sequence of symbols and growth significantly improved, and better than the existing similar detection algorithm. Finally, the detection algorithm can enhance the security capacity of the system to a certain extent. The active mechanism of eavesdropping detection based on subspace method, the signal subspace based method, with the increase of base station the pilot signal matrix dimension and the quantity of samples, the empirical spectral distribution will gradually converge to its limit distribution, it can be the limit matrix The spectral distribution as the theoretical basis to determine the empirical spectral distribution is normal, whether it contains active eavesdropping signal component illegal. Firstly, on Legal signal components for large dimensional random matrix theory the limiting spectral distribution, then calculate the theoretical value of the test statistic. Remove the noise component containing characteristic spectrum in the use of large dimensional random matrix theory, and then the hypothesis test to determine the existence of active eavesdropping. Simulation results show that the detection mechanism of the subspace method can effectively detect the active eavesdropping behavior based on the ratio of the performance of the base station and the legitimate user antenna number of sampling points, the base station receives the signal to noise ratio, increasing the active eavesdropping signal power is significantly enhanced, and better than the similar comparison of algorithms.

【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.3

【相似文献】