摘要
摘要
近年来，为了实现更安全的信息传输和更高的频谱资源利用率，面向全双工(Full-duplex, FD)物理层安全的预编码技术被提出并研究。全双工通信技术可实现同时同频收发信息，基于物理层安全的预编码技术主要利用无线信道的特性实现通信系统的安全性，所以系统的安全性与信道状态信息(Channel State Information, CSI)密切相关。在实际的通信系统中，很多原因都会造成非精确的CSI。在非精确CSI的情况下，非鲁棒性预编码方案很难满足系统的安全性，因此研究非精确CSI 下的鲁棒性预编码方案具有重要意义。本文针对全双工双向安全传输模型，以保密传输为准则，优化设计收发信机结构，探索在双向全双工传输模式下的物理层安全性能极限。具体研究内容如下：
(1) 考虑在精确CSI信道条件下，针对传统预编码方案计算复杂度高的难题，本文提出了两种低复杂度的预编码方案。一种是首先根据子凸问题的对偶问题(Dual Problem, DP)得到半闭合解，然后证明该算法的收敛性；另一种是首先采用迫零(Zero-Forcing, ZF)算法消除自干扰，获得次优的预编码方案。仿真表明，两种全双工预编码方案性能远优于半双工的预编码方案。
(2) 考虑在有界CSI误差模型下，设计具有对抗CSI误差的鲁棒性预编码方案。由于鲁棒保密传输容量的优化问题均是非凸的，不易解决，本文提出了两种鲁棒性预编码方案。当窃听信道误差为确定性有界CS
I误差模型时，和窃听信道协方差的误差为确定性有界CSI误差模型时，本文通过转化含有误差的约束条件和凸差(Difference-of-Convex, DC)规划将非凸问题转为一系列凸优化问题。仿真结果表明，两种鲁棒性预编码算法均满足收敛性；当信道存在误差时，鲁棒性预编码方案优于非鲁棒性预编码方案。
(3) 考虑在随机CSI误差模型下，设计具有对抗随机误差的鲁棒性预编码方案。现有随机误差模型通常假定精确已知误差的分布函数，但在物理层安全应用中，该假设很难满足。本文考虑了一种基于二阶矩的统计CSI误差模型，该模型仅依赖于误差均值和方差，无需已知完整的误差分布函数，因而对误差分布函数具有鲁棒性。针对统计CSI误差模型，本文提出了三种鲁棒性预编码方案。当仅已知窃听信道误差的均值和方差，未知具体分布时，方案一采用马尔科夫(Markov)不等式近似转化概率约束条件，方案二将概率约束条件进行等价的转换。当窃听信道均值和二阶矩存在不确定度，且未知具体分布时，将概率约束条件进行保守的估计。仿真分析结果表明三种算法均具有收敛性和鲁棒性。
关键词：物理层安全，全双工，预编码技术，鲁棒性，凸优化
I
ABSTRACT
In recent years, in order to achieve safer information transmission and higher spectrum resource utili
zation, pre-coding techniques for full-duplex (FD) physical layer curity have been propod and studied. Full-duplex communication technology can realize transmit and recept information at the same time. The precoding technology bad on physical layer curity mainly us the characteristics of wireless channel to realize the curity of communication system. Therefore, the curity of system is cloly related to the channel state information (CSI). In the actual communication system, many reasons will cau imperfect CSI. In the ca of imperfect CSI, nonrobust precoding scheme is difficult to satisfy the curity of the system, so it is very important to study the robust precoding scheme under imperfect CSI. In this thesis, a full-duplex bi-directional cure transmission model is propod to optimize the structure of the transceiver by using the cret transmission as a criterion to explore the physical layer curity performance limit in bidirectional full duplex transmission mode. The specific rearch contents are as follows:
Firstly, for the perfect CSI, aiming at the problem of high complexity of traditional precoding scheme, two low complexity precoding schemes are propod in this thesis. One is to obtain the mi-clod solution according to the dual problem (DP) of the convex subproblem, then prove the convergence of the algorithm. The other is to u the zero-forcing (ZF) algorithm to eliminate the lf-interference and then obtain the suboptimal precoding scheme. Simulation results show that the performance of the two full-duplex precoding schemes is far superior to the half-duplex precoding scheme.
Secondly, for the bound CSI error model, a robust precoding scheme with CSI error is designed. Since the optimization problem of robust confidential transmission capacity is non-convex and difficult to solve, two robust precoding schemes are propod in this thesis. When the eavesdropping channel error is a determined and bounded CSI error model, and the eavesdropping channel covariance error is a determined and bounded CSI error model, this thesis transform the constraint with channel error and u the Difference-of-Convex (DC) programming to convert the nonconvex optimization problem into a ries of convex optimization problems. The simulation results show that the two robust precoding algorithms satisfy the
一如既往英文
II
convergence. When the CSI is imperfect, the robust precoding scheme is superior to the non-robust precoding scheme.
Finally, for the stochastic CSI error model, a robust coding scheme with random error is designed. The existing random error model usually assumes a well-known error distribution function, but in physical layer curity applications, this assumption is difficult to satisfy. This thesis considers a statistical CSI error model bad on the cond moment. The model depends only on the mean and
covariance of the error, and does not need to know the complete error distribution function, so it is robust to the error distribution function. For the statistical CSI error model, three robust precoding schemes are propod in this thesis. When only kown the mean and covariance of the eavesdropping channel error and the exact distribution is not available, the scheme one us the Markov inequality to approximate the probability constraint condition; the scheme two converts the probability constraint condition equivalently. When only kown the estimated first and cond-order moment information on Eve’s CSI and the exact distribution is not available, the probability constraint is conrvatively estimated. The simulation results show that the three algorithms are convergent and robust.
Keywords:physical layer curity, full-duplex, precoding technique, robustness, convex optimization
III
目录
couple第一章绪论 (1)
1.1 引言 (1)
1.2 课题研究背景及意义 (2)
1.2.1 研究背景 (2)
1.2.2 研究意义 (3)
1.3 课题研究现状 (4)
1.4 课题主要工作 (5)
1.5 本论文的结构安排 (6)
第二章物理层安全及全双工通信相关技术介绍 (8)
2.1 引言 (8)danny moder
2.2 全双工通信技术 (9)
2.2.1 自干扰抑制 (9)
2.2.2信道容量 (10)
400岁的格陵兰鲨鱼
2.2.2.1点对点双向通信 (10)
2.2.2.2中继通信 (11)
2.3 物理层安全技术 (12)
2.3.1 物理层安全的提出 (12)
2.3.2 物理层安全的相关理论 (13)
2.3.3 物理层安全相关技术研究现状 (14)
2.4 本章小结 (15)
第三章精确CSI下的全双工保密通信系统预编码技术 (16)
3.1 系统模型及其相关研究 (16)
3.1.1 相关研究背景 (16)
有志者事竟成 英文
3.1.2 系统模型 (17)
3.2基于FD-ADC的预编码方案 (19)
3.2.1 SSRM问题的重构 (19)
3.2.2 FD-ADC算法 (20)
3.2.3 降低FD-ADC算法复杂度 (21)
3.2.4 FD-ADC算法收敛性 (24)
3.3 基于FD-ZF的预编码方案 (25)
IV
3.4 仿真结果与分析 (28)
3.5 本章小结 (31)
第四章有界CSI误差下的全双工保密通信系统预编码技术 (32)
4.1 引言 (32)
4.2 信道误差有界的全双工物理层安全鲁棒性传输方案 (33)
4.2.1 系统模型 (33)
4.2.2 基于S-lemma的鲁棒性预编码方案 (34)
4.2.3 仿真结果与分析 (37)
大学专业排行榜4.3信道协方差误差有界的全双工物理层安全鲁棒性传输方案 (39)
4.3.1 系统模型 (39)
4.3.2 基于DP的鲁棒性预编码方案 (41)
4.3.3 仿真结果与分析 (43)
progress是什么意思4.4 本章小结 (45)
accessible第五章统计CSI误差下的全双工保密通信系统预编码技术 (46)
5.1 引言 (46)
5.2已知CSI误差均值和方差的全双工物理层安全鲁棒性传输方案 (47)
5.2.1 系统模型 (47)
5.2.2 基于Markov不等式的鲁棒性预编码方案 (48)
5.2.3 基于等价变换的鲁棒性预编码方案 (50)
5.2.4 仿真结果与分析 (53)
深圳英语培训机构5.3 已知部分CSI均值和二阶矩的全双工物理层安全鲁棒性传输方案 (59)法语歌
5.3.1 系统模型 (59)
5.3.2 基于保守估计的鲁棒性预编码方案 (60)
5.3.3 仿真结果与分析 (66)
5.4 本章小结 (69)
第六章全文总结与展望 (70)
6.1 全文总结 (70)
6.2 后续工作展望 (71)
致谢 (72)
参考文献 (73)
攻读硕士学位期间取得的成果 (80)
V
