Date & Time:

Wednesday, July 13, 2022 @ 9:20-11:20

Location:

TBD

Organizers:

Maurice Heemels, Dragan Nesic, Alejandro I. Maass, Romain Postoyan.

Abstract:

Motivated by scenarios where the communication or the computation resources are limited, event-triggered control consists in smartly generating transmissions between the plant and its controller according to the system state to achieve desirable stability and performance properties, while only “communicating” when necessary. This alternative to traditional time-triggered (periodic) sampling has been a very active field of research within the control community for more than a decade and various results and techniques are now available. It feels time to review the state of the art, looking back, and identifying interesting open problems, thereby looking forward. As such, the first objective of this tutorial session is to review the basic concepts underlying many of the existing event-triggered control techniques. We mostly concentrate for this purpose on the stabilization problem of a single dynamical system. Afterwards, we will show how hybrid systems theoretical tools are instrumental to model, analyze, design and interpret event-triggered controllers in a unified way. The last part of the session will be dedicated to recent advanced results as well as to the validation of event-triggered control on experimental set-ups.

The tutorial session will consist of the following talks:

• Basic concepts and techniques in event-triggered control (Maurice Hemeels)
• Hybrid modeling and small-gain interpretation of event-triggered control (Dragan Nesic, Alejandro I. Maass)
• Advanced results on event-triggered control (Romain Postoyan)

Date & Time:

Thursday, Jule 14, 2022 @ 9:20-11:20

Location:

TBD

Organizers:

Geethu Joseph, Chandra R Murthy

Abstract:

This tutorial presents a overview of the classical control theoretical concepts through the lens of compressed sensing. We revisit the concepts of observability, controllability, and stabilizability of a discrete-time linear dynamical system with sparse initial states or inputs. The sparse initial states naturally arise in systems like an epidemic spreading, and sparse control inputs model limited system resources such as energy, bandwidth, or network constraints. This tutorial has two parts. The first part focuses on the observability of a sparse initial state of the system, and the second part examines the controllability and stabilizability of the system using sparse controls. In the first part, we describe the use of compressed sensing algorithms to estimate a sparse initial state and develop rigorous sufficient conditions to be satisfied by the system to be observable. We show that systems that are unobservable using classical control theory can become observable when the underlying sparsity is exploited. In the second part, we present simple rank-based conditions for testing the controllability and stabilizability of linear systems using sparse inputs. We then analyze the concrete example of a budget-constrained external agent controlling the opinion of a social network. We also discuss efficient algorithms for designing sparse control inputs exploiting the special sparsity structures in the system. Overall, the tutorial provides a comprehensive overview, critical insights, and deep discussion of recent theoretical and algorithmic advances related to sparsity-constrained linear dynamical systems.

The tutorial session will consist of the following talks:

• Introduction to Compressed Sensing and Sparsity in Linear Dynamical Systems (Chandra R Murthy)
• Estimation of Sparse Initial State in Linear Dynamical Systems(Chandra R Murthy)
• Observability of a Linear Dynamical System with Sparse Initial State (Geethu Joseph)
• Controllability of a Linear Dynamical System Using Sparse Inputs (Geethu Joseph)
• Design of Sparse Control Inputs for Linear Dynamical Systems (Chandra R Murthy)

Date & Time:

Thursday, July 14, 2022 @ 13:30-15:30

Location:

TBD

Organizers:

Alex van Delft, Silvia Mastellone, Tariq Samad.

Abstract:

This tutorial focuses on the gap between control research and industrial practice. It is based on a paper recently published in “Control Engineering Practice” (https://doi.org/10.1016/ j.conengprac. 2021.104737) “The impact of control research on industrial innovation: What would it take to make it happen?” co-authored by Prof. Dr. Silvia Mastellone and Dr. Ir. Alex van Delft. In this paper a framework is proposed to close the gap between fundamental control research and practice towards catalyzing technology innovation. The paper highlights key drivers for innovation, limitations encountered in practice, and suggested research directions. Based on this framework, the authors are preparing an edited volume, to be submitted to the book series on “Control Systems: Theory and Applications” of IEEE Press. This tutorial will create insights and engagement about the urgency to close the gap between academia and industry, and will generate insights on the research demands per industry cluster.

The tutorial session will consist of the following talks:

• Introduction (Tariq Samad)
• Introduction of concepts: control research and the impact on industrial innovation (Alex van Delft, Silvia Mastellone)
• Case study: process industry (Alex van Delft)
• Case study: energy & power conversion (Silvia Mastellone)
• Interactive wrap up (Alex van Delft, Silvia Mastellone)

Date & Time:

Friday, July 15, 2022 @ 9:20-11:20

Location:

TBD

Organizers:

Thomas L. Chaffey, Rodolphe J. Sepulchre.

Abstract:

Graphical analysis of feedback systems has a long and rich history in control theory. Even in the age of state-space modeling and algorithmic analysis, it remains a tool of choice for instructors, students, and practitioners. This tutorial will revisit the topic in the light of the novel concept of the scaled relative graph (SRG). The SRG provides a concept of magnitude and phase for the incremental analysis of a nonlinear operator, very much aligned with the classical frequency-domain characterisation of an LTI operator. The tutorial will provide an introduction to the concept, discuss its significance for nonlinear system analysis and its relationship to classical graphical tools such as the Nyquist curve of a LTI system. Several case studies will be presented where the SRG-based graphical analysis outperforms state-of- the art system analysis. The tutorial is intended for students and practitioners with a basic background in graphical system analysis (Nyquist criterion, circle criterion, describing function analysis) as well as to researchers interested in the many open questions underlying the numerical approximation of SRGs.

The tutorial session will consist of the following talks:

• Introduction – graphical tools in nonlinear control (Rodolphe J. Sepulchre)
• Graphical algebra of Scaled Relative Graphs (Thomas L. Chaffey)
• SRGs and incremental stability (Thomas L. Chaffey)
• Conclusion – ongoing research and open questions (Rodolphe J. Sepulchre)

Date & Time:

Friday, July 15, 2022 @ 13:30-15:30

Location:

TBD

Organizers:

Vincent Andrieu, Pauline Bernard, Michelangelo Bin, Lucas Brivadis, Florent Di Meglio.

Abstract:

The objective of this tutorial is to give an overview of the current theory of Kazantzis-Karavis/Luenberger observers or nonlinear Luenberger observer. We thus bring together the various theoretical results that have been obtained over the past 20 years. The objective is also to show that these approaches have an important impact in different fields of control theory. Then, we show that these approaches have a great interest in the application domain. Finally, we show the different open problems that surround these methods.

The tutorial session will consist of the following talks:

• Theory of Kazantzis-Kravaris/Luenberger observer (Vincent Andrieu, Lucas Brivadis)
• Application of KKL observer: electrical motor (Pauline Bernard)
• KKL observers in output regulation (Michelangelo Bin)
• Machine learning and KKL observers (Florent Di Meglio)
• Deep KKL as a tool for prediction and further development (Vincent Andrieu)