Prof. Jérôme Cieslak , University of Bordeaux, France
Title: Hybrid health-aware supervisory control framework with a prognostic decision-making
Jérôme Cieslak received a MSc in Electronics, Electrotechnical and Automatics from University of Bordeaux 1 in 2004. He received his PhD degree in automatic control from University of Bordeaux 1 in 2007. Since 2008, he is an associate professor of control engineering with the University of Bordeaux, France. His expertise areas include Fault Diagnosis and Isolation (FDI), Fault-Tolerant Control (FTC) and intelligent strategies for complex aerial systems (aircraft, satellite, UAV, …). Especially, the transient management due to the interaction between the FDI and FTC units is his topic of great interest. More recently, the development of a robust artificial pancreas based on biosensor measurements has been investigated to type 1 diabetic subjects. He received the authorization to conduct researches (HDR) in 2019. He has published 26 articles published in international journals indexed with impact factors, more than 50 communications, 1 scientific book, and 4 book chapters. He is a co-holder of 13 patents in the aerospace field. He was involved to several aerospace projects at the European level as the FP7-ADDSAFE project (2009-2012) on fault detection and diagnosis for aeronautical applications, the European GARTEUR FMAG(16) on fault-tolerant control. Recently, he is involved in the national project DIABLO granted by the French National Agency for Research (ANR).
In the area of health management, the research on Fault Detection and Isolation (FDI) has firstly received growing attention since the first patent in 1970. The concept of fault-tolerant computer emerged in 1970s has been progressively extended by the FDI community to introduce the research field of Fault Tolerant Control (FTC). FTC systems aim at guaranteeing that a process keeps fulfilling its mission even in the presence of failures, although possibly in a degraded mode. Despite the evident interaction between FDI and FTC algorithms, the research on FDI and FTC methods has often evolved separately. Typically, it was usually assumed in the FTC literature that a perfect FDI information is available (i.e. no detection delay, …). Unfortunately, this inherent imperfection can irreparably affect stability and performance levels. This makes the need of reconfigurable strategies that explicitly take into account the imperfections between FDI and FTC parts of great importance. To address this weakness, supervisory control concept has been recently applied to FTC purpose. Based on dwell-time conditions, this solution seems particularly efficient since it has been proved that input/output stability is guaranteed even if the FDI unit fails, during a short period, to identify the correct faulty operating mode. However, reconfiguration transient still exists and can corrupt the mission. This makes the need of enhanced FTC strategies that can forecast a fault occurrence to perform a switch, in a more appropriated moment, of great importance. Motivated by the works developed by NASA in 2010s, the design of an enhanced active FTC scheme based on the concept of supervisory control is addressed for uncertain Linear Time Invariant (LTI) systems. The improvement lies on the introduction of a Prognostic Decision Making (PDM) unit and a Virtual Fault Mechanism (VFM) in the supervisory FTC setup. The main purpose is to forecast the switching control thanks to the fault occurrence prediction to better mitigate the reconfiguration transients, while input/output stability is guaranteed by the dwell-time conditions. The proposed hybrid health-aware control framework gives thus an architecture where diagnosis, prognostic and accommodation tasks can work together in harmony. This work is applied to an academic aeronautical benchmark to highlight its principle.
Prof. Andrzej Dzieliński , Warsaw University of Technology, Poland
Title: Dynamic systems modelling using fractional order calculus
Graduated with MSc in Electrical Engineering in 1983, Department of Electrical Engineering, Warsaw University of Technology, Warsaw Poland. PhD in Electrical Engineering in 1992, WUT. DSc in Electrical Engineering in 2002, WUT. Works at the Institute of Control and Industrial Electronics, from 1983, recently as a full professor. Deputy Director of the Institute for Education 2005-2008, Deputy Director of the Institute for Research 2008-2012, Director of the Institute from 1.09.2012
Ecole des Mines de Paris, Fontainebleau, France, 1.09.1990 – 30.09.1990, visiting researcher
University di Roma „La Sapienza”, Rome, Italy 1.02.1991 – 31.07.1991, research associate
University of Glasgow, Glasgow, UK 1.04.1994 – 31.03.1997, research fellow,
Pennsylvania State University, State College, USA 1.06.1999 – 30.06.1999, visiting scientist,
San Diego State University, San Diego, USA 1.02.2002 – 28.02.2002, adjunct professor,
University of Aalborg, Aalborg, Denmark 1.09.2004 – 15.09.2004, visiting professor,
Nicolaus Copernicus University, Torun, Poland 1.10.2003-29.02.2004 and from 1.03.2011, visiting professor,
Author and co-author of over 140 scientific papers in automatic control of 2-D systems, adaptive control, nonlinear control, neural networks, fractional order systems, image processing, modelling and simulation, control education published in journals, books and conference proceedings. Co-author of the manual Foundations of Control Theory (in Polish Podstawy Teorii Sterowania’’, WNT, Warszawa, 4 editions: 2005, 2007, 2009, 2013).
Ten times awarded WUT Rector’s Award for research and education achievements. Polish Academy of Science Award in Technical Sciences in 2004. Polish Minister of Education Award in 2005.
Scopus H index = 12, citations number: 1023.
Participated in many research grants of 3, 4, 5, 6 and 7 Framework Programmes of Scientific Research of European Union, EU research programmes COST and EUREKA, EU higher education programmes TEMPUS and Leonardo da Vinci and National Science Foundation (USA).
Associate Editor of Journal of Circuits, Systems and Signal Processing. Member of IFAC Technical Committee on Control Education, IEEE – senior member. Member of Control and Robotics Committee of Polish Academy of Sciences
Supervisor of 7 completed PhD theses, supervisor of over 40 MSc theses. Reviewer of 18 PhD and 7 DSc theses.
In this talk an overview of recent results of modelling and identification of some dynamic systems, behaviour of which can be better grasped
when using fractional order calculus, are presented. Such systems include ultracapacitors, thermal processes, viscoelastic matherials and many others.
Models derived this way may be of better use for some control algorithms e.g. of the process of charging and discharging in ultracapacitors.
Brief introduction to fractional system modelling is given. Some models of selected physical systems of fractional (non-integer) order, which is the consequence
of specific internal structure of the system under consideration, are introduced.
Theoretical considerations on modelling methods and new models derivations are given.
Models properties are discussed and are verified by simulations and real live experiments.
The models presented justify some well known behaviours of the elements and circuits where they are applied.
Dr Philippe Goupil, Airbus, Aircraft Control,
Title: Industry 4.0: Challenges and opportunities for fault detection and
diagnosis in avionic systems
Philippe Goupil received the PhD degree in signal processing from the National
Polytechnic Institute, Toulouse, France. He works at the AIRBUS design office in Toulouse as an expert in fault/anomaly detection for avionic systems. He is also in charge of advanced fault detection and diagnosis R&T activities dedicated to real-time industrial applications. In particular, he works on model-based and data-driven approaches. He has been involved in the European GARTEUR Flight Mechanics Action Group FM-AG16 (2004-2008, www.garteur.org) on Fault Tolerant Control (FTC) and in the French project SIRASAS which dealt with innovative and robust strategies for spacecraft autonomy (2007-2010). He was the AIRBUS representative in the FP7 European Project ADDSAFE (2009-2012) which focused on Advanced Fault Detection and Diagnosis towards a more Sustainable Flight Guidance and Control (http://addsafe.deimos-space.com/) and in the FP7 European Project RECONFIGURE (2013-2016) which dealt with aircraft GNC technologies that facilitate the automated handling of off-nominal events. He is the author or co-author of 22 industrial international patents and of about 70 conference or journal articles. He has been the industrial supervisor of five PhD students. Philippe Goupil is a member of the International Federation of Automatic Control (IFAC) Technical Committee on Aerospace, the IFAC Technical Committee Safeprocess and the IFAC Industry Committee. He served many times as IPC member and reviewer for several conferences and journals. He gave 5 plenary talks during international conferences. He co-authored, with A. Zolghadri, D. Henry, J. Cieslak and D. Efimov, the book Fault Diagnosis and Fault Tolerant Control and Guidance for Aerospace Vehicles, published by Springer.
The Guidance, Navigation and Control function is a key element of a flying
system like an aircraft. It relies especially on avionic systems which
consist of all electric, electronic and computing components embedded
on-board an aircraft. Most of these systems, like e.g. the Flight Control
System, are critical systems in the sense that any abnormal behavior can
impact the vehicle performances, control and trajectory. In this context,
a Fault Detection and Diagnosis (FDD) module is required to detect,
isolate and possibly estimate any unusual conditions.
The term Industry 4.0 commonly refers to the fourth industrial revolution
and generally focuses on smart and digital factories equipped with
augmented means like wireless connectivity and sensors, advanced
automation, cyber physical systems, internet of things, etc. with the
ultimate goal to optimize the production rate and costs, to improve
manufacturing quality and to enhance safety just to name a few. Another
aspect of Industry 4.0 is that it mainly relies on big data collection
processed by data analytics technics like e.g. machine learning,
statistical analysis, anomaly detection.
In this new era where more and more data is available, computational
capacities are still growing and some advanced technics are mature, this
talk will focus on the opportunities and the associated challenges of the
Industry 4.0 for the FDD field in avionic systems.
This talk will start by a brief overview of classical model-based and
data-driven technics used in avionics systems, where engineering knowledge
is generally key. Then, it will propose some potential opportunities like
e.g. the improvement of state-of-the-art methods, predictive maintenance,
aircraft-centric monitoring, modelling, Verification & Validation, design
tuning and maturation. And finally it will broach the associated
challenges like e.g.: what about the quality and representativeness of the
data ? Is real-time learning really feasible on-board an avionic system ?
How to explain the results in view of the certification of an avionic
All these challenges and opportunities will be illustrated with concrete