Dissertation Defence: Advancing Security in Large-scale Systems under False-data-injection Cyberattacks

Hanieh Tabatabaei, supervised by Dr. Ahmad Al-Dabbagh, will defend their dissertation titled “Advancing Security in Large-scale Systems under False-data-injection Cyberattacks” in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Electrical Engineering.

An abstract for Hanieh Tabatabaei’s dissertation is included below.

Examinations are open to all members of the campus community as well as the general public. Please email ahmad.aldabbagh@ubc.ca to receive the Zoom link for this exam.

Abstract 

The integration of communication, computation, and control in networked systems-such as power, transportation, and water infrastructures-has improved efficiency and scalability while also increasing exposure to adversarial interference. In cyber-physical systems, the tight coupling between cyber and physical layers implies that cyber threats can directly affect physical operation. Reported incidents demonstrate that false-data-injection cyberattacks can seriously compromise system integrity by corrupting communicated signals, particularly when they remain stealthy to conventional detection schemes.

Focusing on such cyberattacks, and motivated by these challenges, this dissertation investigates observer-based cybersecurity distributed frameworks for large-scale systems, where subsystems are coupled through distributed physical and communication interconnections. The dissertation focuses on three main contributions.

First, a secure estimation framework is developed for states, outputs, and injected attack signals enabling detection and isolation, and secure control to asymptotically recover original system performance, under cyberattacks targeting communicated sensor measurements between subsystems. Under non-stealthy cyberattacks, full signal reconstruction is achieved, whereas under stealthy cyberattacks, partial or full state reconstruction is obtained depending on the system interconnection topology.

Second, in scenarios where partial reconstruction is achievable under covert cyberattacks, a detection and isolation framework is first developed to identify compromised communication channels, allowing each subsystem to distinguish between manipulated and secure information. Building on this, a reconstruction framework is proposed that, under suitable conditions on the physical interconnection structure, guarantees full signal reconstruction.

Third, the attacker’s perspective is adopted to identify structural vulnerabilities in nonlinear large-scale systems to provide insights for design of more effective and resilient defence mechanisms, under cyberattacks on communication between the process and the controller. Necessary and sufficient conditions for covertness are derived on the system and detector, and the resources required by a knowledgeable yet resource-constrained attacker, demonstrating that achieving covertness in nonlinear systems is generally more challenging than in linear settings. Four classes of covert cyberattacks are then characterized based on the sufficient conditions on the attacker’s knowledge and access to communicated data.

The proposed frameworks are validated through simulation studies inspired by real-world power-system applications.