Dipartimento di Ingegneria informatica, automatica e gestionale

 

 

Through the last years, discrete-time dynamical systems have been finding a renewed and increasing interest. This is mainly due to their power of modeling a wide range of phenomena linked to modern and in-fashion technologies, including opinion dynamics and social networks, machine learning algorithms for artificial intelligence, big-data event-based analysis or cyber-physical systems, to cite a few. 

Discrete-time behaviors are indeed ubiquitous nowadays in several systems, where the dynamics is purely intermittent or hybrid, as characterized by the co-existence of both continuous-time (flow) and discrete-time (jump) components. In this perspective, several methodological problems arise from the applications. Most of them can be traced to the energy and dissipation properties of complex discrete-time systems that are composed of the interconnection of elementary components through nonlinear elements. Accordingly, a clear understanding of those properties for purely discrete-time systems is necessary. However, if such understanding is advanced in the continuous-time framework, the discrete-time counterpart is still underinvestigated with a lot of unanswered questions. 

The objective of this seminar is to make a small overview on some recent results in energy-based modeling and control of discrete-time systems with particular emphasis on dissipation and passivity. The classes of systems under study are those that admit a port-Hamiltonian representation. The corresponding structures and properties, to date, are still unclear in the discrete-time context despite the increasing interest.  In this sense, a formal characterization of the discrete-time port-Hamiltonian representations will be presented. Then, on this basis, new control methodologies are illustrated to solve stabilization and regulation problems for classes of dynamical (discrete-time or digital systems). Finally, the way these results help in formulating new discrete-time consensus algorithms for enforcing consensus on dynamical networks (involving discrete behaviors) is shown. Those algorithms, based on a suitable energy-based interpretation of the information flow, are shown to overcome several issues that are intrinsic to the ones that are usually employed in practice. 

 

 

Mattia Mattioni received his Bachelor Degree (Laurea) and Master of Science (Laurea Magistrale) in Control Engineering from La Sapienza, Università di Roma (Italy), both Magna cum Laude. In 2015, he received the double degree for the Master de Recherche en Automatique, Traitement du Signal et des Images (M2R ATSI)  through the bilateral agreement Elisa with Université Paris-Sud. 

He received the PhD in System and Control Theory in 2018 from La Sapienza, Università di Roma (Italy) and Universitè Paris-Sud (France) through a joint program with DIAG A. Ruberti (La Sapienza, Università di Roma) under the supervision of Salvatore Monaco and Dorothée Normand-Cyrot. In 2019, he received the SIDRA Award for the best Italian PhD thesis in Automatica; he was a finalist for the Prix des Meilleures Thèses du GDR MACS et de la Section Automatique du Club EEA (France). From October 2018 to August 2020, he was a post-doctoral fellow (assegnista di ricerca) at DIAG "A. Ruberti" (La Sapienza, Università di Roma). Since August 2022, he has been Assistant Professor (Tenure Track, RTDb) at DIAG "A. Ruberti" (La Sapienza, Università di Roma). Since January 2024, he's been Associate Editor for the IEEE Control Systems Letters. He's been member as Associate Editor of the Conference Editorial Board of the IEEE Control Systems Society since 2022 and Corresponding Member of the IFAC Technical Committee on Non-linear Control Systems since 2020. 

His research is mainly concerned with hybrid systems, networked systems and nonlinear systems under sampling and time-delays.