Beamforming For Physical Layer Security

With the wide application of wireless communication technology, the importance of ensuring the security of wireless communications is self-evident. The traditional encryption approach can provide security for wireless system by the secret key and the encryption algorithms. However, the dynamic topology of mobile networks makes it difficult to ensure encryption key not be got by illegitimate users and eavesdroppers’ capacity of cracking encryption algorithm is increasing due to the development of computational science. These two problems have been chronic illness of the traditional encryption approach. Different from the traditional encryption approach, physical layer security aims at exploiting the characteristics of the physical layer, such as fading and noise, to ensure the security of the wireless communication system. Therefore, physical layer security is more reliable and has drawn more and more attention.Consequently, as a technology which plays a key role in whether the multi-antenna systems and collaborative communication systems can work on a high secrecy rate, beamforming technology naturally becomes a hot topic. From the perspective of the physical layer security, beamforming designs for each class of multi-antenna systems and collaborative communication systems can be divided into two categories:one aims at maximizing the system’s communication efficiency while avoid eavesdropping, i.e. secrecy capacity based; another one tries to weaken eavesdropping while ensure communication efficiency, i.e. signal-to-noise ratio based.In this thesis, we study beamforming design for MISO system and decode-and-forward (DF)-based collaborative system from the angle of secrecy capacity and signal-to-noise ratio (SNR). Totally, the main work can be summarized as follows:1. Under the assumption that the channel state information (CSI) about eavesdropper is not perfect, we make a secrecy-capacity-based robust beamforming design for a MISO system which is constituted of a multi-antenna source, a single-antenna destination and a single-antenna eavesdropper. Through semidefinite programming (SDP), we approximate the optimization design problem as a convex optimization problem to obtain the solution of the original design problem. The simulation result show that, our design performs better than the non-robust design in terms of the worst-case secrecy rate.2. Under the assumption that the CSI about eavesdropper is not perfect, we make a SNR based robust beamforming design for a MISO collaborative system which is worked under cooperative jamming scheme. In order to minimize the SNR at the eavesdropper while guarantee the SNR at the legitimate receiver, we jointly optimize the beamforming vector and the spacial covariance of artifical noise. In the process of solving the optimization desin problem, we first approximate it as an SDP problem by neglecting the rank-one constraint. Because the rank of solutions of the SDP problem is not one, we cannot get the optimal solution of the original problem through it. So we formulate a new optimization problem whose optimal solution is also the optimal solution of the SDP problem and must be rank-one. Through decomposing the optimal solution of this new problem, the optimal solution of the original design problem can be got. Proofs and simulation results justify our method.3. With individual power constraint, we make a secrecy capacity based beamforming design for a DF based collaborative system which comprises a single-antenna source, a single-antenna destination, a single-antenna eavesdropper and some single-antenna relays. We assume that perfect CSI is achievable and then propose a method to solve the optimization design problem. Firstly, we focus on solving a subproblem which is got by fixing the denominator of the objective function of the optimization design problem. Then, we change the value of the denominator and obtain the optimal solution of the original design problem through one-demensional search. Subsquently, zero-forcing based simplification is discussed. Simulation result show the performance of the simplified design is similar to the original design.