Secure and energy-efficient wireless networks

Abstract

The underlying theory of physical layer security has considered the fundamental characteristics of a wireless propagation environment and has opened a new domain of research for wireless security. On one hand, physical layer security has been proven to be the strictest form of wireless security. On the other hand, it has operated independently of cryptography- based security at the higher layers. The recent development of physical layer security has shown its compatibility in various wireless systems and its potential in many applications. In this thesis, the secure performance has been investigated for emerging wireless systems, including massive MIMO relaying systems, cell-free massive MIMO systems, energy-harvesting MIMO- OFDM systems, and machine-learning-based authentication systems. The thesis has also proposed security designs in energy-efficient manners.

In order to characterize the behaviours of eavesdroppers in wireless systems, two types of eavesdroppers, namely, passive eavesdroppers and active eavesdroppers, have been modelled. The countermeasures against these two types of eavesdroppers have been developed based on the specific properties of each type. The thesis has resorted to stochastic geometry to model the presence of passive eavesdroppers and evaluated the amount of leaked information on average. Regarding active eavesdropping, the thesis has developed detection mechanisms in order to detect the abnormality in received signals and designed optimal transmission strategies to cope with active eavesdroppers.

In order to maintain the power consumption within a limited budget, the trade-off problems between the level of security and the amount of consumed power have been presented. These trade-off problems have led to lion-convex optimization problems lmt have been solved with the help of path-following algorithms. The existence of at least a sub-optimal solution, which can lie found through a path-following algorithm, has confirmed the feasibility of energy-efficient security at the physical layer.

Date of Award	Dec 2019
Original language	English
Awarding Institution	Queen's University Belfast
Supervisor	Duong Tran Anh (Supervisor)