The Physical Layer Security In Multi-antennas Systems

Security has been widely concerned and recognized as a critical issue in wireless communication networks recently, because the openness of the wire-less medium allows unintended receivers to potentially eavesdrop on the trans-mitted messages. As a promising technique to guarantee the communication security, physical layer security (PLS) technique has attracted increasing atten-tion. The basic idea of PLS is to exploit the physical characteristics of wireless channel and maximize the uncertainty about the source messages at the eaves-dropper.Meanwhile, the growing energy demand of wireless communication sys-tem has drawn attention of both academia and industry. Energy harvesting (EH)has been considered as a promising technique to expand the lifetime of network by providing safe and green energy from the environment. In EH, ambient radio frequency (RF) can be the energy source for energy receivers (ERs). The RF signal can carry information and energy at the same time, as a consequence, the scheme named simultaneous wireless information and power transfer (S WIPT)comes into being. It is also worth noting that ERs are not only able to collect en-ergy, but also intercept the messages sent to other intended users, which brings great challenges to the security of the SWIPT systems.Based on this background, the thesis studies the physical layer security in multi-antenna systems, and proposes secure schemes based on multi-antenna techniques for different practical scenarios. The main innovations and contri-butions of the paper are as follows:1) For the two-user multiple-input single-output interference network with some ERs which may act as eavesdroppers to decode the messages sent to in-tended users, we study the joint beamforming and artificial noise design which can maximize the total energy subject to the individual power at each trans-mitter and the secrecy sum rate requirement. We decompose the non-convex optimization problem into a two-stage problem. To reduce computational com-plexity, we also design a zero-force scheme.2) For a wiretap multi-pair decode-and-forward network (MP-DFRN), our goal is to maximize the secrecy rate of the secure user subject to the transmit power constraints and QoS (Quality of Service) requirements of other users,and minimize the transmit power with the secrecy rate constraint of the secure user and QoS requirements of other users. Aiming to solve the non-convex problem, we proposed an efficient iterative algorithm.3) Considering a more practical scenario, in which the eavesdropper's channel state information (CSI) is imperfect in MP-DFRN, we propose a joint robust beamforming and power allocation design. The channel matrix uncertainty and channel covariance uncertainty models are taken into consid-eration, and we aime to maximize the worst-case secrecy rate of the secure user, subject to the transmit power constraints and QoS requirements of other unclassified users in these two models.4) For a wiretap DF relay network with imperfect CSI in the presence of multiple eavesdroppers, we propose a scheme to maximize the worst-case se-crecy rate in the ellipsoid-bounded model, then, we also investigate the scheme to maximize the lower bound of the secrecy rate of the Gaussian wiretap channel model. Our proposed design constitutes non-convex problems, and we propose an iterative algorithm based on the idea of semi-define relaxation.