Development of Resilient Software System for Cyber-Security Enhancement of Aviation Cyber-Physical Systems
The aviation industries recently have shown an interest in transferring their commercial aircraft into fully connected and network aircraft systems. This desire will lead to transforming aircraft in general and commercial aircraft in specific, into Aviation Cyber-Physical Systems (ACPS). However, like any Cyber-Physical system, ACPS is vulnerable to cyber-attacks that adversaries can mount through the communication network infrastructure. Adversaries can gain unauthorized access to the aircraft’s connected physical devices, such as sensors, actuators, or the control system through the communication network. This dissertation research work developed a security software system for detecting and defending the communication network of ACPS from two major cyber-attacks: Distributed Denial of Service (DDoS) and Fault Data Injection (FDI) Attacks. A novel security strategy (method) was developed to detect and defend against the cyber-attacks mentioned above. The system of interest (SOI) of this research comprises three subsystems: Communication Network Simulator (CNS), Attacks Detection and Classification (ADC), and Defense Against Attacks (DAA). CNS subsystem was developed to simulate dynamics of the communication network based on a discrete-event approach. This allows for establishing an aircraft networked control system (NCS) via a communication network to ensure implementation of the closed-loop control architecture. The CNS was developed using the SimEvents MATLAB/Simulink toolbox. The developed CNS was tested and validated on the selected aircraft associated with an adaptive control system, which produced the expected results. ADC subsystem was developed to detect DDoS and FDI attacks of the communication network of ACPS based on the signal’s classification approach. In the ADC, an Artificial Immune System (AIS) algorithm was developed and implemented to detect and drop malicious communication packets of the aircraft network traffic produced by both cyber-attacks. The detection accuracy of the AIS algorithm reached 96% based on true positive and true negative detection rates. The DAA subsystem was developed and implemented to defend the communication network of the ACPS against both DDoS and FDI cyber-attacks. A nonlinear autoregressive neural network with external inputs (NARX) was developed and used to reconstruct or estimate the network dropped packets. The estimation accuracy of the NARX reached 0.99, using the correlation coefficient (R-value) metric.
Computer Engineering|Artificial intelligence
Abdulaziz A Alsulami,
"Development of Resilient Software System for Cyber-Security Enhancement of Aviation Cyber-Physical Systems"
ETD Collection for Tennessee State University.