Transmit Design Optimization for Wireless Physical Layer Security

Security is one of the most important issues in communications. Conventional techniques for achieving confidentiality in communication networks are based on cryptographic encryption. However, for wireless networks, this technique is faced with more challenges due to the open nature of the wireless medium as well as the dynamic topology of mobile networks. In the 1970's, Wyner proposed a physical layer-based approach to achieve perfectly secure communication without using encryption. One of the key problems of Wyner's approach is how to optimally design the transmit signal such that a high secrecy rate (i.e., the data rate at which the confidential information can be securely transmitted) can be achieved. In our work, we aim to solve this transmit signal optimization problem under various scenarios using convex optimization techniques. Specifically, the thesis consists of the following three main parts:;In the first part, we consider a multi-input single-output (MISO) scenario, where a multi-antenna transmitter sends confidential information to a single-antenna legitimate receiver, in the presence of multiple multi-antenna eavesdroppers. Our goal is to maximize an achievable secrecy rate by appropriately designing the transmit signal. The challenge of this secrecy rate maximization (SRM) problem is that it is a nonconvex optimization problem by nature. We show, by convex relaxation, that this seemingly nonconvex SRM problem admits a convex equivalent under both perfect and imperfect channel state information (CSI) cases. Our result also indicates that transmit beamforming is an optimal transmit strategy, irrespective of the number of eavesdroppers and the number of antennas employed by each eavesdropper. This provides a useful design guideline for practical implementations.;In the second part, we consider a scenario where the transmitter is able to simultaneously generate artificial noise (AN) to interfere the eavesdroppers during the transmission of the confidential message. While the efficacy of AN in improving the system security has been demonstrated in many existing works, how to jointly optimize the AN and the transmit signal is still an unsolved problem. In this part, we solve this AN-aided SRM problem under the same scenario as the first part, and give an efficient, semidefinite program (SDP)-based line search approach to obtain an optimal transmit signal and AN design under both perfect and imperfect-CSI situations.;In the last part, we consider a secrecy capacity maximization (SCM) problem for a multi-input multi-output (MIMO) scenario, where the legitimate receiver and the eavesdropper are both equipped with multiple antennas. This MIMO-SCM problem is a generalization of the previous MISO-SRM problems. So far there is no known efficient algorithm to solve this problem in a global optimal manner. Herein, we propose an alternating optimization algorithm to tackle the SCM problem. The proposed algorithm has a nice iterative water-filling interpretation and is guaranteed to converge to a stationary solution of the MIMO-SCM problem. Extensions to robust SCM are also investigated in this part.;Besides the above three main results, this thesis also developed some approximate solutions to the multicast SRM of multiple MISO legitimate channels overheard by multiple MIMO eavesdroppers, and to the outage-constrained SRM of an MISO legitimate channel overheard by multiple MISO eavesdroppers.