Physical Layer Security for Securing Resource-Constrained Networks

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

Recent developments in society are trending towards the use of large networks comprising a large number of small devices. Traditional cryptographic techniques are no longer an attractive solution due to excessive memory and power requirements. Physical Layer Security (PLS) promises to be a good fit for forthcoming networks due to its potential for relatively low computational complexity. However, being a relatively new field, there are challenges to be addressed before PLS is integrated into real-world systems. After establishing the underlying theory in information theory and wireless communications, this thesis covers the background of PLS, identifies challenges, and suggests novel solutions for confidentiality, and authentication attacks in short-range systems. All suggested methods for security consider one end of the legitimate communication link to be severely constrained in terms of computational capabilities. The work presented after the literature review (chapter 4) challenges one of the most common assumptions made in key-based PLS, namely, the half-wavelength channel decorrelation assumption. It proves that such an assumption brings secrecy vulnerabilities which are brought to attention and quantified. The results motivate the definition of secure distance which facilitates the quantification of secrecy performance in terms of spatial channel correlation. Observations on the outcomes of chapter 4 motivate the solutions provided in the next two chapters. Chapter 5 presents a method that exploits base-station cooperation for the secure transmission of small data such as keys. The method addresses the main challenge of keyless PLS which is the requirement of a positive secrecy gap. Although spatial channel correlation is treated as an impairment for key generation purposes in chapter 4, chapter 6 demonstrates how it can be used to our benefit for protecting against authentication attacks in short-range systems via two novel methods. The first method targets relay attacks and replay attacks - the two most common impersonation attacks in short-range communications. The second method aims to verify the distance of devices that communicate through backscattering modulation and are equipped with little or no local power. The final research chapter finds a solution to the problem of a high reconciliation cost in key-based PLS. More precisely, chapter 7 presents a key agreement protocol which, different to existing key agreement protocols, achieves an arbitrarily low key disagreement rate without increasing the computational complexity. The practicality of the suggested key agreement protocol is successfully tested on a series of typical IoT boards.
Date of Award3 Oct 2023
Original languageEnglish
Awarding Institution
  • University of Bristol
SupervisorOliver T Johnson (Supervisor) & Angela Doufexi (Supervisor)

Cite this

'