The development of novel physical layer security algorithms to mitigate cognitive radio attacks
thesisposted on 2022-03-28, 02:12 authored by Sasa Maric
Since the implementation of the first public-access networks, attackers have looked to take advantage of vulnerabilities in network security to gain an unfair advantage. In recent times, wireless networks have increasingly been integrated in our everyday lives. Science-fiction style automated homes and societies have increasingly become a reality. Today's wireless devices possess high cognitive ability, they dynamically adjust according to their environment and user preferences to ensure maximum comfort for their users. As a result, a global network of interconnected wireless devices has been growing exponentially for the past few decades. Previous radio-frequency spectrum allocation has failed to predict this growth, which has resulted in extreme congestion in some bands and low utilisation of others. Cognitive Radio, a collection of intelligent methods, is seen as the most promising solution. To increase effciency they allow secondary users (users that do not have a regulatory right to use a frequency channel) to utilise allocated frequency bands when they are not being utilised by paying users (primary users). However, cognitive radio implementation has been delayed several times because of its susceptibility to a number of security attacks, specifically in the physical layer. As such, a taxonomy of new attacks has been identfied, which could not be mitigated by standard security algorithms that were developed for conventional wireless networks. The primary aim of this thesis is to mitigate the effects of physical layer attacks in Cognitive Radio Networks (CRN). In particular, two attacks have been identified as the most serious threats to cognitive radio security. These are a Primary User Emulation Attack (PUEA), which involves an attacker emulating the properties of primary users in order to gain an unfair advantage over other secondary users and a Spectrum Sensing Data Falsification Attack (SSDFA), during which an attacker intentionally manipulates messages containing spectrum sensing information in order to trick secondary users into miss-diagnosing the status of a primary user. In this thesis, we present a number of algorithms to combat the vast array of attacks within the physical layer. In particular we present a number of novel, highly effective, low computational complexity algorithms that can be implemented to completely eradicate these attacks and render them ineffective. Since many of the devices that make up a cognitive radio network have battery and computational complexity constrains, our objective was to develop mitigation algorithms that they are lightweight and can be implemented effectively.