Study On Physical Layer Security Of Wireless Networks

Security is a key issue of wireless networks, and it is the basic condition on which awireless network can be popularized and applied successfully. Moreover, as the quantity ofusers grows and network environment gets worse, the security problem becomes more andmore serious. How the security performance can be assured and enhanced, has always beenstudied by researchers. Similar with wired networks, traditional wireless network securityadopts the key-based cryptography to enhance the security performance, and all kinds ofencryption algorithms have been proposed. However, when a new encryption algorithm isproposed, a corresponding attacking algorithm will always appear. Therefore, even though thekey-based cryptography can get a good security performance, it will pay a high price forcomputational complexity.Unlike the key-based cryptography, physical layer security deals with security problemsfrom a perspective of information theory. The openness of physical channel, multiple pathsand fading, all these characteristics make wireless networks’ security problem more seriousthan wired network. However, some disadvantageous characteristics could be utilized from adifferent angle. Physical layer security solves the security problem by utilizing thesecharacteristics.This dissertation studies physical layer security of some kinds of wireless networks, andproposes corresponding alrithoms to enhance their security performances. The contributionsof this dissertation include:1) Artificial noise based physical layer security of MISO cognitive radio networks isstudied from the QoS perspective, and the transmitter is designed by optimizing thebeamforming vector and noise convariance together. By utilizing semi-definite relaxingtechnique, the complicated optimization problem is converted into a semi-definiteprogramming, which can be easily solved. Furthermore, a robust transmitter is designed byusing worst-case optimization approach on the premise of knowing the region of uncertaintiesof CSI. Therefore, the security performance can still be assured even with CSI errors.2) The secure communication in MISO cognitive radio networks via cooperativejamming is studied. A cooperative node is utilized to transmit artificial noise, so as to interferewith illegal user’s wiretap. For dealing with the complex optimization problem, twosuboptimal schemes are proposed: complete orthogonal projection (COP) and partialorthogonal projection (POP). In these schemes, the original optimization problem isdecoupled into two sub-problems, which solve the beamforming vector and power allocation, respectively. Simulation results show that both COP and POP could enhance the secrecy rateefficiently. The COP algorithm is with low computational complexity, but it is practicableonly when N h≥3. The POP algorithm can get a better performance than the COP algorithmand without any condition on the transmitter’s antenna. However, it’s more complexed thanthe COP algorithm.3) Physical layer security of AF cooperative relay networks is studied. For dealing withthe information leakage in the first phase, a new cooperative jamming method is proposedbased on analog network coding (ANC) technique. A fraction of power is allocated to thedestination node for transmitting a jamming signal in the first phase. According to ANC, thejamming signal could be emitted by the destination node and will not interfere itself. Asuboptimal solution is obtained by solving beamforming and power allocation separately.Numerical results show that the proposed method can restrict the information leakageefficiently, and the security performance is enhanced apparently.