Online Seminar Series for IFAC TC 1.2 Adaptive & Learning Systems

The main objectives of our "Online Seminar Series for IFAC TC 1.2 Adaptive & Learning Systems" are:

• Promote the latest research results in the field of Adaptive and Learning Systems;
• Create a forum for high-quality discussion of both theoretical and practical perspectives;
• Reach and engage the researchers and practitioners in our community worldwide;
• Provide a platform for the young generation to network with the rest of the world.


Organization of the Seminar Series:

We plan to run the online seminar series on a monthly basis. We will invite speakers at different career stages, both within and outside of the TC, to share their research with the community. To encourage and train the young generation, e.g., PhD students. We will also organise special events within the seminars series to provide a venue for them to share and discuss their research with their peers and experts in the community.

The seminar will be open to all researchers and practitioners across the world. It will be scheduled at a suitable time (in the afternoon, UTC time) to maximise the possible attendance. The lectures will be recorded and hosted on IFAC YouTube Channel (upon the permissions of the speakers and approval of IFAC).

Details of Next Seminar: 

The next webinar in this series will be given by Professor Florian Dörfler from Swiss Federal Institute of Technology (ETH).

Title: Direct Adaptive Learning of the LQR
Time: May 02, 2024 (05:00pm CET)

Abstract: The linear quadratic regulator (LQR) problem is a cornerstone of automatic control, and it has been widely studied in the data-driven setting. In the first part of the talk, we show how to bridge different problem formulations and propose a novel, direct, and regularized version of the LQR. We start from indirect certainty-equivalence LQR, i.e., least-square identification of state-space matrices followed by a nominal model-based design, formalized as a bi-level program. We show how to transform this problem into a single- level, regularized, and direct data-driven control formulation, where the regularizer accounts for the least-square data fitting criterion. For this novel formulation we carry out a robustness and performance analysis in presence of bounded noise. In the second part of the talk, we propose an adaptive method to learn this solution. By adaptive, we mean an online method using closed-loop data, in a non-episodic fashion, and with recursive algorithmic implementation. Our approach is based on a covariance parameterization of the direct, data-driven, and regularized LQR and an explicit calculation of the policy gradient using a batch of persistently exciting data. We establish the global convergence of our method via a projected gradient dominance property in presence of bounded noise. Finally, all our theoretical results are validated with simulations and demonstrate the computational and sample efficiency of our method.

To join the seminar, you may use the meeting link below:

Join Microsoft Teams Meeting
https://msteams.link/S3JN
Meeting ID: 321 669 544 041
Passcode: pZpCkj

We will open the meeting 15 minutes earlier. The participants must join earlier - 10 minutes before the start. It will take some time to let them all in :-)

All of you are welcome! See you there!

Tiago Roux Oliveira (Chair of the TC 1.2: Adaptive and Learning Systems)
Bing Chu  (Vice-Chair for Social Media of the TC 1.2: Adaptive and Learning Systems)

Past Webinars:

  • Professor Anuradha Annaswamy from MIT, USA.
    • Title: The role of adaptation in learning, safety, and optimality
    • Time: April 25, 2024 (09:30am ET)
    • Video: TBC
  • Professor Romeo Ortega, Instituto Tecnológico Autónomo de México.
    • Title: New Robust Parameter Estimators and Systems Reparameterizations: Dealing with Lack of Excitation and Nonlinear Parameterizations
    • Time: September 26, 2023 (11am CDT, 4pm UTC)
    • Video: https://youtu.be/N3qfiakYsCI 
  • Professor Denis Dochain, Université catholique de Louvain.
    • Title: Automatic Control and Biological Systems: a long quiet river?
    • Time: April 12, 2023 (9am CDT, 2pm UTC) 
    • Video: https://youtu.be/VZ5y7HckfeM
  • Professor Na Li, Harvard University, Cambridge.
    • Title: Scalable Distributed Control and Learning of Networked Dynamical Systems
    • Time: February 15, 2023 (7am PST, 9am CDT, 3pm UTC) 
    • Video: https://youtu.be/sGOzBUWTcBk
  • Professor Magnus Egerstedt, University of California, Irvine.
    • Title: Constraint-Based Control Design for Assured and Long-Duration Autonomy
    • Time: January 11, 2023 (9am PST, 11am CDT, 5pm UTC)
    • Video: https://youtu.be/ATqcna2YoDg 
  • Professor John Ringwood, Maynooth University. 
    • Title: Energy maximising control for wave energy systems: an extremum seeking problem?
    • Time: December 19, 2022 (4pm UTC)
    • Video: https://youtu.be/vQCyfawXVog 
  • Professor Martin Guay, Queen’s University. 
    • Title: Data driven control of unknown nonlinear systems using extremum seeking control
    • Time: May 27, 2022 (8am PST, 10am CDT, 3pm UTC)
    • Video: https://youtu.be/MMr8M9JpGD8
  • Professor Tamer Basar, University of Illinois Urbana-Champaign.
    • Title: Policy Optimization for Optimal Control with Guarantees of Robustness
    • Time: April 27, 2022 (8am PST, 10am CDT, 3pm UTC)
    • Video: https://youtu.be/7sjv24wNyBA
  • Professor Miroslav Krstic, University of California, San Diego.
    • Title: The Magical Worlds of Adaptive Stabilization and Optimization
    • Time: March 14, 2022 (8am Pacific Daylight Time, 3pm Universal Time UTC)
    • Video: https://youtu.be/X_N-0N22VZE

https://tc.ifac-control.org/1/2