Guest talk by Prof. Riccardo M.G. Ferrari
Critical infrastructures such as power grids, communication and energy networks; potentially dangerous industrial processes such as nuclear or chemical plants; transportation systems or autonomous robots. These are examples of systems for which safety and resiliency should be an integral part of their design. The occurrence of anomalies such as faults or cyber attacks can lead to unacceptable losses, or simply make the operation of those systems uneconomical. For instance, in the offshore wind energy sector Operation & Maintenance costs can already reach up to 30% of the total lifetime cost. Detecting and accommodating faults before they lead to extreme consequences is thus a key requirement. Model-based techniques constitute a powerful way to detect anomalies, by comparing the predictions of a mathematical model (or digital twin as often referred to in the industry) of the system to real-time measurements coming from sensors. The central problem then becomes detecting when differences between the two are caused by physiological uncertainties present in the model and the sensors, and when they are due to an actual, pathological fault. From this point of view, thus, uncertainties are a challenge that must be fought. What this talk will show, someway surprisingly, is that uncertainty can also cover another role, this time beneficial. This is the case of methods aimed at guaranteeing security, and thus safety, of cyber-physical systems from malicious cyber-attacks. In the presence of a malicious attacker, a defender can consciously inject uncertainty in the data that the attacker may likely use to prepare an attack, thus misleading them. This is the case for instance in Differential Privacy mechanisms, watermarking techniques or, when uncertainty is brought to an extreme case, encryption algorithms. The talk will start by stating the fundamental problems in anomaly detection and fault tolerant control, how robustness against uncertainty can be guaranteed, and then move to later results in cyber-security where uncertainty-augmenting methods are used to the benefit of defenders.
Riccardo M.G. Ferrari is a Marie Curie Alumnus and an Associate Professor in Fault Tolerant Control in the Delft Center for Systems and Control, within the Faculty of Mechanical, Maritime, and Materials Engineering (3mE) at Delft University of Technology, The Netherlands. He investigates how to make dynamical systems safe and resilient against faults, malicious cyber-attacks and degradation phenomena, fighting uncertainty while doing so. His research is applied to problems in wind energy, in the aerospace and in the automotive sectors, in particular for electric and for cooperative autonomous vehicles. He received the “Laurea” degree (Cum Laude and printing honours) in Electronic Engineering in 2004 and the Ph.D. degree in Information Engineering in 2009, both from the University of Trieste, Italy. He is the co-recipient of the 2023 Hugo Schuck Award from the American Automatic Control Council, and has authored and co-authored over 90 papers published in international peer-reviewed journals and conference proceedings. He covered a unique career path that led him to hold both academical and industrial R&D positions, in particular as researcher and executive manager in the field of process instrumentation and control for the steelmaking sector. In 2005 he earned a B.A. in Classical Piano from the “G. Tartini” Conservatory of Music of Trieste (Italy).