Download files
Abstract
This thesis considers the use of location information as a means to enhance physical layer security in wireless communication systems. The thesis has four main contributions.
First, we consider a very general system model for wiretap channels, namely the case of Rician fading as described by a Rician K-factor. Under such fading, the channels can vary from pure line-of-sight to pure Rayleigh. For all values of Rician K-factors, we obtain the optimal beamformer solution at the source that minimizes the secrecy outage probability, under the assumption that both the channel state information (CSI) from the legitimate receiver, and the location of the eavesdropper, are available at the source. We also consider the impact a friendly jammer can have on the beamformer solution.
Second, we consider a more specific wiretap channel in which a relay is present, and the source is allowed to transmit artificial noise signals to confuse the eavesdropper. Under this more complex scenario we consider the case of pure Rayleigh fading, developing a framework to characterize the secure transmission probability and the effective secrecy throughput of the combined source-relay-destination channel. We then introduce a location-based secure transmission scheme that maximizes this throughput. In this scheme, the two key inputs are the CSI from authorized transceivers, and the location of the eavesdropper. We also investigate in detail the impact of the uncertainty in the eavesdropper's location on the secrecy performance, showing how our proposed scheme can still allow for secrecy when only a noisy estimate of the eavesdropper's location is available.
Third, we further investigate the secrecy performance of the relay wiretap channel where the location of the eavesdropper is described by a Poisson point process. In this new model we allow both the source and the relay to transmit artificial noise signals. We then develop an artificial-noise-aided secure transmission scheme that maximizes the secrecy throughput, while satisfying a secrecy outage probability constraint.
Fourth, to see how location information on the eavesdropper can have the same role as the information on the eavesdropper's CSI, we look at a case where no information whatsoever on the location of the eavesdropper is available, but imperfect CSI on the eavesdropper is available. Reverting back to a no-relay case, we examine again the secrecy performance of a wiretap channel assuming imperfect knowledge of the eavesdropper's CSI. We develop three beamforming schemes, applicable for different settings at the source, which maximize the achievable secrecy rate. We discuss how the introduction of location information can lead to similar solutions when no eavesdropper's CSI is available, thus demonstrating a form of equivalence between these different information sources.
The work reported in this thesis provides insights into the design of new secure transmission schemes based on location information. Such insights should assist in the development of new physical layer security solutions over a wide range of emerging wireless communication systems.