Plenary Sessions

The Organizing Committee is pleased to host Andrés Marcos, Alisa Rupenyan and Henk van Waarde as keynote speakers at the 43rd Benelux Meeting on Systems and Control.

Mini-course I — Andrés Marcos

Biography

Prof. Andrés Marcos has 20+ years of work experience on robust control for aerospace systems, about half in academic posts and half in industry. During his career he has been principal investigator for over 22 projects, directly responsible for obtaining 17M€, and has published over 140 peer-reviewed conference and journal publications. Together with his team he has flight tested robust controllers and fault detection and isolation / fault tolerant schemes in manned aircraft (DLR’s ATTAS and JAXA’s MuPAL-alpha), autonomous Vertical Take-Off and Landing vehicles (DLR’s EAGLE), and small test rockets (NDUT).

Currently, he is the “Beatriz Galindo Distinguished Senior Investigator”, a 5-year personal talent-attraction award by the Spanish Government, at the Universidad Carlos III de Madrid (UC3M) where he leads the UC3M-SENER Aerospace chair and is the director of the Master in Space Engineering (MISE) and the Center for Satellite Research and Technology (CSAT). He is also the founder and scientific director of Technology for Aerospace Control Ltd (TASC), a UK-based SME specialized in bridging the gap between academia and industry for aerospace control systems.

Tentative title of the mini-course

Launcher modeling, control, and analysis: from academic teaching to industrial transfer

Abstract

Despite the elegant and established theory behind robust control, as well as the numerous applications across different fields, there always has been strong reticence by industry in adopting robust design methods. The main problem has been that the large gap between the theoretical developments performed by academia and the practical implementation requirements demanded by industry, which coupled with the emphasis of the latter on legacy knowledge, made very difficult to change their established design processes. In addition, from a human resources perspective, in general engineering graduates are not exposed to robust control and even if they are, it is still not widespread that they are in senior technical positions. Nonetheless, since the mid-90s the robust control community (academia and industry) have demonstrated that these approaches are specifically apt for application to aeronautical and space systems, and further, with the advent of powerful fixed-structure control design methods in the 2010s  this has led to a not so quiet revolution. For space, it can be claimed (and shown) that many of the telecommunication satellites flying around Earth as well as the most advanced future observation missions (LISA, ATHENA) or even launchers (VEGA) are using robust methods and tools.

This two-hour plenary mini-course presents the application of robust control modeling, design, and analysis methods to the VEGA launcher ascent control, first from a graduate classical-modern teaching perspective and in the second hour via a real industry technology transfer project. The first part of the talk is oriented towards demonstrating that the fundamental concepts used in control must be reconciled with a physical understanding in order to train the future generation of control engineers. And the second, to demonstrate that robust approaches offer a methodological and incremental design process that leads to more robust and higher performance designs for complex, industrial-level aerospace systems. The talk concludes with an overview of robust design aerospace applications that included flight tests.

 

Plenary lecture — Alisa Rupenyan

Biography

Dr. Alisa Rupenyan holds the endowed professorship in Industrial AI from the Johann Jakob Rieter foundation at the Zurich University for Applied Sciences and specializes in continuous optimization and automation of industrial systems. Previously, she was a PI and senior scientist at the Automatic Control Laboratory at ETH Zurich and at the same time group leader for Automation at Inspire, the technology transfer institute at ETH Zurich. She has a PhD in Physics from Vrije University Amsterdam, and MSc and BSc degrees in Engineering Physics from Sofia University.

She was granted an ETH excellence postdoctoral fellowship to fund her research in the field of high-harmonic spectroscopy for the study of attosecond molecular dynamics. After that, she switched to the industry and lead a team in a Swiss robotics startup combining her experience in spectroscopy with machine learning and control. Her research interests include autonomous machines, decision-making in industrial settings, learning-based optimization and control. She is an innovation expert for the Swiss Innovation Agency, member of the executive committee at the IFAC industry committee, and startup advisor. She is an active member of the NCCR Automation research competence center in Switzerland, leading the efforts in automation for advanced manufacturing.

Tentative title of the lecture

Data-driven optimization and control for advanced manufacturing

Abstract

Data availability and advances in robotics are driving forces for innovations in advanced manufacturing, requiring adaptive control and optimization methods for enhanced functionality and new economic models. The distinct challenges of advanced manufacturing - process variability, drifts, and lack of in-situ measurements - demand approaches that are data-efficient, conform to physical process limits, and prioritize safety. In this talk I will discuss data-driven, predictive, and learning-based optimization methods which enhance productivity, quality, and energy consumption, while respecting safety and operational constraints and making the most out of the available data. I will highlight Bayesian optimization algorithms for process control, approaches for repetitive control incorporating data from the system to accommodate for process changes and drifts, and integrating digital twins in such algorithms to increase productivity on a machine level, as well as on a system level, while making optimal use of the available resources. These algorithms are inspired by real-world applications including additive manufacturing, plasma spray coating, and high-precision motion stages, and I will showcase their application on such systems.

 

Mini-course II — Henk van Waarde

Biography

Dr. Henk van Waarde is an assistant professor in the Bernoulli Institute for Mathematics, Computer Science and Artificial Intelligence at the University of Groningen in The Netherlands. During 2020-2021 he was a postdoctoral researcher, first at the University of Cambridge, UK, and later at ETH Zürich, Switzerland. He obtained the Ph.D. degree (cum laude) in Applied Mathematics from the University of Groningen in 2020. He was also a visiting researcher at the University of Washington, Seattle in 2019-2020.

His research interests include data-driven control, system identification and identifiability, networks of dynamical systems, and robust and optimal control. He was a recipient of the 2021 IEEE Control Systems Letters Outstanding Paper Award, and the 2021 EECI Ph.D. Thesis Award.

Tentative title of the mini-course

Mini-course on data-driven control

Abstract

Advances in sensing technology have led to a massive increase in available data. There is an enormous potential for using these data to guide the design of controllers for a wide range of applications, including power systems, robotics and biological systems. However, before such controllers can be employed in practice, a number of fundamental issues need to be addressed. Indeed, how can we make sure that data-driven control design comes with the same stability and performance guarantees traditionally associated with model-based control? And how to develop data-driven control algorithms that are able to cope with big data sets?

The purpose of this two-part mini-course is to provide an overview of some of the main ideas in data-driven analysis and control. The mini-course is theoretical in nature and the material will be presented in a tutorial style.

In Part 1, I will begin with some historical perspectives, going back to the literature on persistency of excitation and the fundamental lemma by Willems and coauthors. We will see how this literature gave rise to a number of so-called direct data-driven control methods, i.e., approaches that map data directly into control policies without the intermediate step of system identification. Thereafter, I will introduce the notion of informative data as a general framework for studying a variety of data-driven analysis and control problems. The framework will be applied to obtain data-driven Hautus tests for controllability analysis, and conditions for data-driven stabilization using exact data.

Part 2 provides a deepening of the theory developed in the first part of the mini-course. A central question of this part is how to handle data that are corrupted by noise. After a discussion on noise models, I will present several results related to quadratic matrix inequalities and matrix S-lemmas that will pave the way for the design of controllers that are robust to noise. We will consider controllers with stability and performance guarantees in different settings, including the cases of input-state and input-output data.