الفهرس 
Chapter 1: IntroductionOnly 14 pages are availabe for public view
 |  | from | 146 |  |  |
| | | from | 146 | | |
Abstract
The urgency of cybersecurity and cyber resilience in the aviation sector cannot be ignored, and this fact has been highlighted time and again due to subsequent cyber-attacks prior and during the COVID-19 pandemic that denigrate the victim(s) and also lead to political and economic controversies. The importance of cybersecurity has recently increased with the transition of aviation systems from ground-based systems to space-based systems to accept the growth in air traffic capacity. Furthermore, the majority of aviation systems have unsecured designs that make those systems prone to security issues. Additionally, in addition to the catastrophic impact of COVID-19 on the global economy in general, and on aviation in particular, the pandemic has showed increased levels of cyber-attacks.
Therefore, the ability to have a methodology for cybersecurity risk assessment at aviation domain is an important issue in the recognition of potential threats, and assessment of the likelihood and risk levels, such that risks can be reduced to tolerable levels by applying the suitable mitigation control procedures. To achieve this, this dissertation composed of three main parts. Firstly, we develop a detailed framework for a methodology for aviation cybersecurity risk assessment. Secondly, the proposed methodology is applied on various aviation systems, including air to ground communication, navigation aids, surveillance, and system-wide information management (SWIM). The proposed methodology is also applied to recognize the most common increased cyber-attacks accompanied with the outbreak of COVID-19 and their risk levels. Finally, we address the cybersecurity in Automatic Dependent Surveillance-Broadcast (ADS-B). ADS-B is considered as the backbone of next-generation air traffic surveillance systems owing to its low cost and high accuracy. To check the verification of ADS-B data of aircraft, multilateration (MLAT) techniques have been proposed. MLAT achieves low accuracy in determining aircraft locations. Recently, a new technique utilizing a theoretically calculated Time Difference of Arrival (TDoA) fingerprint map has been presented. This methodology depends less on the sensor deployment and has improved accuracy than MLAT. However, this technique is not sufficiently precise for estimating aircraft positions and also needs a long processing time. Unlike other techniques, this dissertation introduces a precision surveillance utilizing an Actual TDoA-Based Augmentation System (ATBAS). This technique utilizes historically stored data from the OpenSky network to train the proposed TDoA fingerprint map.
Regarding applying our proposed methodology for cybersecurity risk assessment on various aviation infrastructure systems, our analysis shows that in communication systems, the system with the topmost risk level is very-high frequency voice communication while the system with the smallest risk level is controller-pilot data link communication. For navigation, satellite-based navigation is the navigation system with the topmost risk level while the systems with smallest risk level are ground radio navigation aids. For surveillance, ADS-B is that with the topmost risk level while that with the smallest risk level is traditional secondary surveillance radar. Additionally, the risk level of attacks in SWIM is medium. Furthermore, our analysis shows that the common increased cyber-attacks associated with the outbreak of COVID-19 in aviation are phishing and ransomware. The likelihood of phishing and ransomware attacks is very likely during the pandemic than before it while the impact on flight operations is low, and the subsequent risk level is tolerable during the pandemic. Additionally, we propose a set of possible mitigation measures to those cyber-attacks.
Regarding applying our proposed cybersecurity techniques in ADS-B, our results shows that the precision of our ATBAS technique in determining the aircraft positions is greatly better than those of the expected TDoA and MLAT systems by 48.86% and 56.93%, respectively. Additionally, our technique reduced the processing time by 77% in comparison with the expected TDoA system.