A Research Of Robust Transmit Technology On Physical Layer Security

The security of confidential message becomes even more important in wireless networks due to the open nature of wireless medium. Traditionally, high-layer key encryption method is used to provide security. But the security of this method is facing a series of challenges due to the improvement of processing ability of computer. Different from the key encryption method, physical layer security(PLS) enhances security by exploiting the physical characteristic of wireless channel, without relying on secret keys.Most of the existing PLS transmit strategies are based on the assumption of perfect channel state information(CSI). But it is difficult to obtain perfect CSI in practical communication scenarios. The performance of these non-robust transmit strategies may not meet the needs of the system. In view of the above questions, this paper investigates the PLS robust transmit strategies with imperfect CSI. Firstly, robust security transmit strategies for the cooperative communication systems are studied. Then, we put forward the robust transmit strategies for broadcast channel with integrated services. The details of our work are as follows:Firstly, a joint cooperative relaying and jamming transmit strategy for secure relay networks is proposed. Specifically, we jointly optimize the relaying weight vector and the artificial noise covariance matrix, aiming at maximizing the worst-case achievable secrecy rate under imperfect CSI of the eavesdroppers. Such an optimization problem,unfortunately, is nonconvex and challenging to solve directly. To overcome this difficulty,we recast it as a two-layer optimization problem, where the inner-level part is handled by the semidefinite relaxation(SDR) technique and the outer-level part is handled by a onedimensional search. Numerical results show that the proposed joint cooperative relaying and jamming strategy always outperform the relaying without jamming strategy.Secondly, a robust transmit strategy with outage constraints for the Multiple Input Single Output(MISO) wiretap channel with a cooperative helper is proposed. Specifically, we design the transmit covariance matrices of the transmitter and the helper, aiming at maximizing the outage constrained achievable secrecy rate under imperfect CSI of the eavesdropper. We reformulate the optimization problem with non-convex outage constraints as a sequence of convex problems. Numerical results show that our proposed transmit strategy always outperforms the ZF strategy.Thirdly, a robust transmit strategy for MISO Gaussian Broadcast Channel with integrated services is proposed. Specifically, we aim to study the robust design of the input covariance matrices of both the confidential message and common message to maximize the worst-case achievable secrecy rate region under imperfect CSI of both users. To this end, we adopt a Quality of Service(Qo S) constrained method, and recast the original nonconvex optimization problem as a sequence of convex ones. Numerical results show that our proposed transmit strategy significantly outperforms the TDMA method. Moreover, we study the rank of the optimal solutions. It proves that, generally, the rank of the optimal input covariance matrix associated with the confidential message is 1, while the rank of the optimal input covariance matrix associated with the common message is less than 4. Further, we prove that for the case of perfect CSI, beamforming is optimal for both common message and confidential message.Finally, a robust transmit strategy for Multiple Input Multiple Output(MIMO) Gaussian Broadcast Channel with integrated services is proposed. Specifically, we jointly design the input transmit covariance matrices of the confidential message and the common message by maximizing the worst-case achievable secrecy rate region under imperfect CSI of the unauthorized user. We reformulate the optimization problem with non-convex constraints and objective function into a semidefinite programming(SDP) problem via the Taylor series expansion. It can be efficiently solved in an iterative fashion with convergence guaranteed. Numerical results show that our proposed transmit strategy significantly outperforms the TDMA method.……...