Séminaire MoFED – 25 novembre 2025 – 14h – Saint-Jérôme
Alessandro Giua, DIEE, Univ. of Cagliari, Italy
Title : State estimation of partially observed discrete event systems under attack
Abstract : Partially observed discrete event systems are a general formalism dating back to the definition of nondeterministic automata. The assumption is that the sequence of events generated by a plant is observed through a mask, so that an agent observing the plant may have incomplete information concerning its evolution and, correspondingly, the past state trajectory and the current state.
The objective of this talk is to describe the basic principles of partially observable discrete event systems showing how the state estimation problem can be addressed for systems subject to cyber-attacks. An operator receives the sensor readings produced by a plant through a communication channel and uses this information to estimate the current state of the plant. The observation may be corrupted by an attacker which can insert and erase some sensor readings with the aim of altering the state estimation of the operator. Furthermore, the attacker wants to remain stealthy, namely the operator should not realize that its observation has been corrupted.
I will show how to determine an automaton, called joint estimator under attack, that describes for each possible observation produced by the plant and for each possible attack, what is the state estimation computed by the operator. Such a structure is obtained by the concurrent composition of two state observers, called attacker observer and operator observer. The joint estimator can be used to determine if there exists a stealthy harmful attack function such that the set of states consistent with the uncorrupted observation computed by the attacker, and the set of states consistent with the corrupted observation computed by the observer, satisfy a given relation.
Biography : Alessandro Giua received a Laurea degree in electrical engineering from the University of Cagliari, Italy, in 1988 and master’s and Ph.D. degrees in computer and systems engineering from the Rensselaer Polytechnic Institute, Troy, NY, USA, in 1990 and 1992, respectively. He is currently Professor of Automatic Control with the Department of Electrical and Electronic Engineering of the University of Cagliari. His research interests include discrete-event systems, hybrid systems, networked control systems, Petri nets and failure diagnosis. He has served as Editor-in-Chief of the IFAC journal Nonlinear Analysis: Hybrid Systems, Senior Editor of the IEEE Trans. on Automatic Control, and Department Editor of the journal Discrete Event Dynamic Systems. He is a Fellow of the IEEE and a Fellow of the IFAC for contributions to Discrete-Event and Hybrid Systems. He is a recipient of the People’s Republic of China Friendship Award.
Cristian Mahulea, Departamento de Informática e Ingeniería de Sistemas, Universidad de Zaragoza, Spain
Title: Efficient Path Planning and Task Allocation for Large Robotic Teams Using Petri Net Structural Properties
Abstract: Coordinating large teams of mobile robots in complex environments remains challenging due to the combinatorial explosion of possible assignments and the need for safe, collision-free motion. This talk presents a new approach that models robot motion and global task specifications using Robot Motion Petri Nets and exploits the total unimodularity of the resulting constraint matrices. By proving this structural property, we relax classically NP-hard integer formulations into tractable linear programs while preserving integrality, enabling solutions that scale to thousands of robots. A two-stage algorithm combines fast LP relaxations with the automatic insertion of minimal synchronization points to guarantee collision avoidance. Comparative evaluations on standard MAPF benchmarks demonstrate significant reductions in computational time and improved scalability over traditional ILP-based or graph-search methods. Applications range from industrial logistics to smart manufacturing and autonomous fleets.
Biography: Cristian Mahulea is Full Professor of Automatic Control at the University of Zaragoza, Spain. He received his M.Sc. in Control Engineering from the “Gheorghe Asachi” Technical University of Iași (Romania) and his Ph.D. in Systems Engineering from the University of Zaragoza. His research focuses on discrete event and hybrid systems, Petri nets, and the planning and control of multi-robot systems, with applications in logistics and healthcare. He serves as Associate Editor of the IEEE Transactions on Automatic Control (TAC), the International Journal of Robotics Research (IJRR), and the Journal of Discrete Event Dynamic Systems (JDES), and has served on editorial boards of several leading journals. Prof. Mahulea has co-authored two books and organized major international conferences in automation.
Yan Monier, LURPA, ENS Paris-Saclay
Title: ADAM: Anomaly Detection by Adaptive Modeling
Abstract: This talk presents ADAM, a project under development that explores a new approach to creating and updating interpretable digital twins for industrial production. Unlike opaque “black-box” AI or costly, heavily hand-crafted models, ADAM leverages existing production data to automatically build and refine digital twins with minimal supervision. It integrates models from discrete event systems, such as automata and Petri nets, while also taking into account time, flows, and dynamic equations to capture the full hybrid behavior of industrial processes. Beyond its methodological advances, ADAM serves as a way to concatenate and operationalize years of research conducted across multiple laboratories into a concrete, real-world problem solver for the industrial field, rather than a purely academic exercise. Its goal is to detect anomalies, failures, and cyberattacks at an early stage while ensuring that the models remain transparent and understandable to operators. Currently under construction and backed by the University of Paris-Saclay, the project aims to validate its potential through pilot collaborations, reduce downtime in real industrial environments, and ultimately contribute to a more resilient, secure, and future-ready manufacturing industry.
Biography: Yan Monier recently completed a PhD in Automation and Control at ENS Paris-Saclay, where he developed methods for hybrid model synthesis that combine physical modeling with data-driven analysis. His recent research included work on cyberattack and anomaly detection in cyberphysical systems and led to the development of software tools applying these methods to realistic laboratory-scale production processes. Focused on making complex industrial systems easier to monitor, predict, and control, Yan now continues this line of research through ADAM, a project that brings interpretable digital twins into real-world production. Passionate about resilience and sustainability, he aims to reduce downtime, strengthen cybersecurity, and enable industries to modernize their processes without costly equipment replacement.