Plenary and Semi-Plenary Sessions
IFAC-V 2020 will feature five Plenary and six Semi-Plenary sessions.
Reminder: All times are Berlin time (CEST, UTC+2h)!
Plenary Sessions
Learning and Control
Plenary Talk: Reflections on the
Learning-to-Control Renaissance (Benjamin Recht)
Panel Discussion: "Control
and learning – is there really a divide?"
Monday, July 13,
13.30-15.00
Summary and
Program
Corona
Talks: Data Science to fight against COVID-19 (Nuria
Oliver), Real-time Global Disease Forecasting for COVID-19 (Sara Del
Valle), Control Strategies for COVID-19/Social Distancing Policies in Brazil (Julio
E. Normey-Rico)
Panel Discussion: "Data Driven Decision Making in the
COVID-19 Pandemic"
Tuesday, July 14, 13.00-14.30
Summary and Program
The Joy of Exploring the Emerging Dynamics of Integrated Systems
Plenary Talk: The Joy of Exploring the Emerging Dynamics of Integrated Systems
(Jing Sun)
Panel Discussion: "Joy or Pain?"
Wednesday, July 15,
13.00-14.30
Summary and
Program
Reinforcement Learning for Process Control and Beyond
Plenary
Talk: Reinforcement Learning for Process Control and Beyond (Jay H. Lee)
Panel
Discussion: "Industrial Potential of Reinforcement Learning"
Thursday,
July 16, 13.00-14.30
Summary and
Program
Robots Interacting with Humans and Unknown Worlds: Pushing the Boundaries of
Nonlinear Control
Plenary Talk: Robots Interacting with Humans and
Unknown Worlds: Pushing the Boundaries of Nonlinear Control (Alin Albu-Schäffer)
Demonstration: "Live Robotics Demos"
Friday, July 13,
13.30-15.00
Summary and
Program
Semi-Plenary Sessions
Integrated Framework of Systems, Data, and Industrial Intelligence towards
Industry 4.0
Plenary Talk: Integrated Framework of Systems, Data, and
Industrial Intelligence towards Industry 4.0 (S. Joe Qin)
Panel Discussion
Tuesday, July 14, 14.30-16.00
Summary and Program
Estimating human objectives from action and interaction
Plenary
Talk: Estimating human objectives from action and interaction (Dana Kulić)
Panel
Discussion: "Control meets neuroscience - modeling human motor behavior via optimal
control"
Tuesday, July 14, 14.30-16.00
Summary and Program
Toward Human Assistive Robotics for Aiding Human Activity
Plenary
Talk: Toward Human Assistive Robotics for Aiding Human Activity (Mihoko Niitsuma)
Panel Discussion: "What are social skills for human assistive robots?"
Wednesday, July 15, 14.30-16.00
Summary and Program
Pushing the limits of performance
Plenary Talk: Pushing the limits
of performance (Maarten Steinbuch)
Panel Discussion: "Can we accelerate the
Impact of Systems and Control?"
Wednesday, July 15, 14.30-16.00
Summary and Program
Going Big: Control Co-Design for Extreme-Scale Wind Turbines
Plenary Talk: Going Big: Control Co-Design for Extreme-Scale Wind Turbines (Lucy
Pao)
Panel Discussion: "Wind turbines with blades longer than a football
field? How is that possible?"
Thursday, July 16, 14.30-16.00
Summary and Program
Safe autonomy with deep learning in the feedback loop
Plenary
Talk: Safe autonomy with deep learning in the feedback loop (George J. Pappas)
Panel Discussion
Thursday, July 16, 14.30-16.00
Summary and Program
Reflections on the Learning-to-Control Renaissance (Benjamin Recht)
Monday, July 13, 13.30-15.00
Session Chair: Frank Allgöwer
Plenary Talk: Reflections on the
Learning-to-Control Renaissance (Benjamin Recht)
Monday, July 13,
13.30-14.20
Bio and Abstract
Panel Discussion: "Control and learning – is there
really a divide?"
Monday, July 13, 14.20-15.00
Abstract: In this panel discussion, we will shed light on different questions, developments, and perspectives related to the interface of Machine Learning and control theory. Among the questions that will be addressed are:
- Will learning-based methods replace established control theory in industrial applications?
- What is missing for learning-based methods to be applicable in safety-critical applications?
- How can we use control theory to improve and robustify machine learning algorithms?
Panelists:
- Carsten Scherer (University of Stuttgart, Germany)
- Thomas Schön (Uppsala University, Sweden)
- Mario Sznaier (Northeastern University, Boston, USA)
- Melanie Zeilinger (ETH Zurich, Switzerland)
The panel discussion is moderated by Frank Allgöwer (University of Stuttgart, Germany).
Corona Session
Please find details on the Corona Session page.
The Joy of Exploring the Emerging Dynamics of Integrated Systems (Jing Sun)
Wednesday, July 15, 13.00-14.30
Session Chair: Hajime Asama
Plenary Talk: The Joy of Exploring the Emerging
Dynamics of Integrated Systems (Jing Sun)
Wednesday, July 15,
13.00-14.05
Bio and Abstract
Panel Discussion: "Joy or Pain?"
Wednesday, July 15, 14.05-14.30
Abstract: A panel of leading researchers, all staying at home in Michigan under COVID-19 special orders, will share their perspectives and inspirations on system integration and emergent dynamics. Robotics, flexible aircraft, manufacturing systems, and transportation infrastructure will all serve as emblematic platforms to showcase interesting dynamic phenomena and grand challenges/opportunities associated with control-oriented system integration, in an attempt to stimulate broad attention and discussion in our control community.
Panelists:
- Kira Barton (University of Michigan, USA)
- Jessy Grizzle (University of Michigan, USA)
- Ilya Kolmanovsky (University of Michigan, USA)
- Huei Peng (University of Michigan, USA)
- Jing Sun (University of Michigan, USA)
Reinforcement Learning for Process Control and Beyond (Jay H. Lee)
Thursday, July 16, 13.00-14.30
Session Chair: Paul Van den Hof
Plenary Talk: Reinforcement Learning for Process
Control and Beyond (Jay H. Lee)
Thursday, July 16,
13.00-13.50
Bio and Abstract
Panel Discussion: "Industrial Potential of
Reinforcement Learning"
Thursday, July 16, 13.50-14.30
Abstract: Reinforcement learning (RL) has come into the limelight following the well-publicized success of Alpha-Go, which demonstrated RL’s potential for solving highly complex, sequential decision-making problems unreachable by traditional optimization approaches. In this panel discussion immediately following the broadcasting of Prof. Jay H. Lee’s plenary talk on “Reinforcement Learning for Process Control and Beyond”, a panel comprising three industrial panelists and two academicians will address the following questions:
- Does RL have potential applications in industry?
- If yes, what should be the primary (or initial) target applications?
- What are the pros and cons of RL over MPC or other established methods?
- Can RL be combined with MPC in a complementary or synergistic way?
Panelists:
- Thomas Badgwell (ExxonMobil)
- Alf Isaksson (ABB)
- Joseph Lu (Honeywell)
- James Rawlings (UCSB)
- Jay H. Lee (KAIST)
The panel discussion is moderated by Paul Van den Hof (Eindhoven University of Technology).
Robots Interacting with Humans and Unknown Worlds: Pushing the Boundaries of Nonlinear Control (Alin Albu-Schäffer)
Friday, July 17, 13.30-15.00
Session Chair: Klaus Janschek
Plenary Talk: Robots Interacting with Humans and
Unknown Worlds: Pushing the Boundaries of Nonlinear Control (Alin
Albu-Schäffer)
Friday, July 17, 13.30-14.40
Bio and Abstract
Demonstration: "Live Robotics Demos"
Friday, July 17, 14.40-15.00
Abstract: During the interactive part of the plenary session, we will have a lab tour with live demonstrations. Three different labs with following topics will present interactive demos:
- Robotics-For Health Care, including the wheeled humanoid robot Justin, the assistive wheelchair EDAN and the teleoperation device HUG
- Legged robotics, including the Humanoid TORO and the quadruped BERT
- Orbital Service Robotics, with focus on control of free-flying systems, consisting of a satellite equipped with a dexterous arm
Integrated Framework of Systems, Data, and Industrial Intelligence towards Industry 4.0 (S. Joe Qin)
Tuesday, July 14, 14.30-16.00
Session Chair: Rolf Findeisen
Plenary Talk: Integrated Framework of Systems, Data, and
Industrial Intelligence towards Industry 4.0 (S. Joe Qin)
Tuesday,
July 14, 14.30-15.15
Bio and Abstract
Panel Discussion: "Ingredients for industrial
intelligence: Machine learning and data analytics vs. systems theory and
modelling"
Tuesday, July 14, 15.15-16.00
Abstract: Machine learning and data analytics are becoming increasingly popular for intelligent production. We discuss the role and interplay of classical system and control theory and modeling in a world of artificial intelligence. What are the pros and cons? Are traditional approaches necessary at all? Where are the limits of machine learning and data analytics? Is it possible to fuse both methods?
Panelists:
- Larry Biegler (Carnegie Mellon University, USA)
- Richard Braatz (Massachusetts Institute of Technology, USA)
- Rolf Findeisen (Otto von Guericke University Magdeburg, Germany)
- Nina Tornhill (Imperial College London, United Kingdom)
- Joe Qin (City University of Hong Kong, China)
- Dimitrios Varvarezos (Aspentech, USA)
Estimating human objectives from action and interaction (Dana Kulić)
Tuesday, July 14, 14.30-16.00
Session Chair: Sandra Hirche
Plenary Talk: Estimating human objectives from
action and interaction (Dana Kulić)
Tuesday, July 14,
14.30-15.25
Bio and Abstract
Panel Discussion: "Control meets neuroscience -
modeling human motor behavior via optimal control"
Tuesday,
July 14, 15.25-16.00
Abstract: In this panel we will cover fundamental questions regarding human sensorimotor control and the application of optimal control ideas to model it. In particular, we will discuss the role of models in recovering human objectives - are they appropriate and do they capture the human body and the central nervous system? Further discussion will concern how to identify the human cost function structure and features and the role of priors, i.e. parametric vs. non-parametric cost functions? Finally, we will discuss the prospect for applications and limitations.
Panelists:
- Dana Kulić (Monash University, Australia)
- Sandra Hirche (Technical University of Munich, Germany)
- Katja Mombaur (University of Waterloo, Canada)
- Jörn Diedrichsen (University of Western Ontario, Canada)
- Martin Giese (University Clinic Tübingen, Germany)
Toward Human Assistive Robotics for Aiding Human Activity (Mihoko Niitsuma)
Wednesday, July 15, 14.30-16.00
Session Chair: Janan Zaytoon
Plenary Talk: Toward Human Assistive Robotics for
Aiding Human Activity (Mihoko Niitsuma)
Wednesday, July 15,
14.30-15.30
Bio and Abstract
Panel Discussion: "What are social skills for human
assistive robots?"
Wednesday, July 15, 15.30-16.00
Abstract: The semi-plenary speaker is working on human assistive robotics for aiding human activity. High functionality and high productivity for robotic agents might be required to assist human activity. However, is it enough if the robot is very functional and very productive to aid human activity? In this panel discussion, let us discuss how we could realize human assistive robotics harmonizing with human life.
Panelists:
- Peter Korondi (Budapest University of Technology and Economics, Hungary)
- Vicente Matellan (University of León, Spain)
- Adam Miklosi (Eötvös Loránd University, Hungary)
- Mihoko Niitsuma (Chuo University, Japan)
Pushing the limits of performance (Maarten Steinbuch)
Wednesday, July 15, 14.30-16.00
Session Chair: Anuradha Annaswamy
Plenary Talk: Pushing the limits of performance
(Maarten Steinbuch)
Wednesday, July 15, 14.30-15.20
Bio and Abstract
Panel Discussion: "Can we accelerate the Impact of
Systems and Control?"
Wednesday, July 15, 15.20-16.00
Abstract: Technology develops at an exponential speed, but is our discipline suitable to track this and have impact? Should we re-balance between theory and practice? Is the next generation of university more linked to industry, and how can the local ecosystems benefit from our capability to cope with complex systems? Does Covid-19 reshape our academic future?
Panelists:
- Maarten Steinbuch (Eindhoven University of Technology, Netherlands)
- Danny Abramovitch (Agilent Technologies, USA)
- Kira Barton (University of Michigan, USA)
- Angelika Peer (Free University of Bozen-Bolzano, Italy)
The panel discussion is moderated by Anuradha Annaswamy (MIT, USA).
Going Big: Control Co-Design for Extreme-Scale Wind Turbines (Lucy Pao)
Thursday, July 16, 14.30-16.00
Session Chair: Jacquelien Scherpen
Plenary Talk: Going Big: Control Co-Design for
Extreme-Scale Wind Turbines (Lucy Pao)
Thursday, July 16,
14.30-15.30
Bio and Abstract
Panel Discussion: "Wind turbines with blades longer
than a football field? How is that possible?"
Thursday, July
16, 15.30-16.00
Abstract: Wind turbines are being designed at ever larger scales in order to decrease the cost of wind energy. But how large can we continue to increase the size of wind turbines? Come find out from a panel of aerodynamic, structural, and control experts who are working as a team to design and control wind turbines at scales ranging from 13 MW to 50 MW.
Panelists:
- Gavin Ananda (Zipline)
- Christopher Bay (University of Colorado Boulder Alumnus)
- Rogier Blom (Controls & Optimization GE Research)
- Mayank Chetan (University of Texas at Dallas)
- Rick Damiani (Colorado School of Mines)
- Alejandra Escalera Mendoza (University of Texas at Dallas)
- Lee Jay Fingersh (University of Colorado Boulder Alumnus)
- D. Todd Griffith (University of Texas at Dallas)
- Kathryn Johnson (Colorado School of Mines)
- Nick Johnson (Colorado School of Mines Alumnus)
- Lucy Pao (University of Colorado Boulder)
- Mandar Phadnis (University of Colorado Boulder)
Safe autonomy with deep learning in the feedback loop (George J. Pappas)
Thursday, July 16, 14.30-16.00
Session Chair: John Lygeros
Plenary Talk: Safe autonomy with deep learning in
the feedback loop (George J. Pappas)
Thursday, July 16,
14.30-15.40
Bio and Abstract
Panel Discussion: "Feedback in the age of
data"
Thursday, July 16, 15.40-16.00
Abstract: The panel discussion will delve into the topics of digitalization and Cyber-Physical Systems and the role of automatic control in this context. In the popular press these broad terms are often associated with data science and machine learning, as these are the disciplines that allow information technologies to make sense of the physical world. Automation and control complement this view and provide the mechanisms that allow information technology to make a difference to the physical world. This aspect of “closing the loop” in the age of data will form the core of the discussion: How can one build on the knowledge gained about the physical world through data science to make decisions that are fed back to the physical world and steer its evolution towards desirable outcomes?
Panelists:
- George Pappas (University of Pennsylvania, USA)
- Anders Rantzer (Lund University, Sweden)
- Claire Tomlin (U.C. Berkeley, USA)
The panel discussion is moderated by John Lygeros (ETH Zurich, Switzerland).