Learning to control: history and challenges of direct data-driven design (Prof. Simone Formentin, Politecnico di Milano)
Systems & Control Seminar
Abstract
In the 1990’s, the Russian scientist Vladimir Vapnik used to say "When solving a problem of interest, do not solve a more general problem as an intermediate step. Try to get the answer that you really need but not a more general one." In machine learning, this is among the key guidelines to solve any estimation problem. In automatic control, engineers often spend most of their time to devise a good model of the system, so as to use it as a basis for controller design. In particular, it is estimated that 75% of the time devoted to a control project is dedicated to modeling. In the era of big data, where huge datasets are available in many applications, it is reasonable to foster a change of rationale, so that data are directly mapped onto control laws, without going through a costly and time-consuming modeling step. In this seminar, we will review the preliminary attempts towards the so-called "direct data-driven design" of feedback controllers and then outline the most effective learning-based control techniques. The most challenging open problems in the field will be finally discussed to lay the ground for future research.
Biographical information
Simone Formentin was born in Legnano, Italy, in 1984. He received his B.Sc. and M.Sc. degrees cum laude in Automation and Control Engineering from Politecnico di Milano, Italy, in 2006 and 2008, respectively. In 2012, he obtained his Ph.D. degree cum laude in Information Technology within a joint program between Politecnico di Milano and Johannes Kepler University of Linz, Austria. After that, he held two postdoctoral appointments at the Swiss Federal Institute of Technology of Lausanne (EPFL), Switzerland and the University of Bergamo, Italy, respectively. Since 2014, he has been with Politecnico di Milano, first as an assistant professor, then as an associate professor. He is the chair of the IEEE TC on System Identification and Adaptive Control, the social media representative of the IFAC TC on Robust Control and a member of the IFAC TC on Modelling, Identification and Signal Processing. He is an Associate Editor of Automatica and the European Journal of Control. His research interests include system identification and data-driven control with a focus on automotive and financial applications.
Termin
18. Jan. 202316:00 - 17:00
