Dear DISC member,
Welcome to the first newsletter of 2019! Next week will start with two new DISC courses and in March we are organizing the DISC Winter Course on Reinforcement Learning. It will be held from March 27 till 29 at Delft University of Technology. A great opportunity to learn more about this topic. You can read more about that further on in the newsletter. Of course the Benelux Meeting is also coming up. Hopefully many of you will join!
Henk Nijmeijer,
scientific director
Martha Otte,
DISC secretariat
In January 2019 the following classes will start:
Mathematical Models of Systems
J. Polderman, K. Camlibel and H. Trentelman
Input Design and Parameter Estimation for Non-linear Systems
H. Stigter and K. Keesman
In February we will offer:
Nonlinear Control Systems
B. Jayawardhana & B. Besselink
For the full list of courses see: https://disc.tudelft.nl/education/course-program/2018-2019-course-program/
Register on the DISC course platform.
The courses are taught at cursus en vergadercentrum Domstad in Utrecht, location Koningsbergerstraat 9. For more information see their website.
This course provides an introduction to reinforcement learning, starting from the basics and ending up with advanced topics like deep RL. The course targets graduate students in control or artificial intelligence/machine learning.
The first afternoon holds an optional session on Markov decision processes, their optimal solution, and basic dynamic programming techniques to find this solution: value and policy iteration (for students who do not have these techniques in their background).
Then, the first full day of the course opens with a discussion of forward-search planning methods for seeking near-optimal sequences of actions. The methods work like receding-horizon predictive control, and exploit insights from tree search in AI and bandit theory in reinforcement learning. We explain how planning can be applied and adapted to switched systems and networked control systems. The afternoon of the second day provides an introduction to the basics of RL including temporal-difference, Q-learning and SARSA methods and actor-critic algorithms. The exploration exploitation trade off will be discussed. Approaches to speed up the learning such as shaping, demonstrations and advice will be introduced.
In the second and final day of the course, algorithms are extended with function approximation techniques, in order to make them applicable to continuous-state and action control, as well as to large-scale discrete-variable problems. We begin by introducing function approximation in general, and then apply it to some basic dynamic programming and reinforcement learning techniques. We then pay attention to policy gradient techniques, as they are a natural way of using continuous actions which are essential in control. After a detour on deep learning, we will discuss some recent Deep RL approaches. Extensions to multi-agent and multi-criteria RL will also be briefly introduced.
Content
•Introduction and Course Overview
•Markov Decision Processes
•Dynamic Programming
•Forward-Search Planning with Applications to Nonlinear Control
•RL Basics
•Exploration Exploitation Trade-off
•Basics of Function Approximation For Continuous States and Actions
•Detour: Deep Learning, Neural Nets and Convnets
•Deep Reinforcement Learning: Value Based
•Deep Reinforcement Learning: Policy Based
•RL extensions : multi-agent RL and multi-criteria RL
•The future and open problems
Lecturers
Anne Nowé is professor at the Artificial Intelligence Lab of the Free University of Brussels, Belgium. In 2019 she is also guest-professor at Cognitive Robotics department of the Delft University of Technology.
Lucian Busoniu is professor at the Automation Department of the Technical University of Cluj-Napoca, Romania.
Prerequisites
This course assumes the students are familiar with the basics of Markov Decision Processes and Dynamic Programming. Students who are lacking this background can attend an introduction to these concepts on Wednesday afternoon.
Date and Location
The course will take place from Wednesday March 27 until Friday March 29, 2019. Wednesday is an optional session introducing Markov decision processes, their optimal solution, and basic dynamic programming techniques to find this solution: value and policy iteration. Students who are familiar with these concepts need not attend.
The location is Delft University of Technology, the Pulse Building.
Registration and fee
Registration fee for taking or auditing a full course is € 250. This fee is waived for DISC members. The registration form is available on the DISC course platform, or send an email to secr@disc.tudelft.nl.
Please register before March 7, 2019.
Credits
You can obtain 1 ECTS for attending the DISC Winter Course. Please note that you have to be present at all sessions on Thursday and Friday in order to obtain the credits. We are discussing the possibility to create a “take home” exam by which you can upgrade to 3 ECTS. At this point this plan is tentative.
The organizers have the pleasure of inviting you to participate in the 38th Benelux Meeting on Systems and Control. The meeting will be held from March 19 – 21, 2019 in “Center Parcs De Vossemeren”, Lommel, Belgium,
The aim of this meeting is to promote research activities and cooperation between researchers in Systems and Control. It is the thirty-eighth in a series of annual conferences that are held alternately in Belgium and The Netherlands. The meeting is organized by KU Leuven.
Abstract submission opens: December 16, 2018
Deadline abstracts submission: January 26, 2019
You will find more information on the website:
https://www.beneluxmeeting.nl/2019/ which is continuously updated.
The 15th IFAC Symposium on Large Scale Complex Systems (IFAC LSS 2019) will be held in Delft on May 26-28, 2019. Tthe plenary speakers are:
Giancarlo Ferrari Trecate (EPFL Lausanne)
Jacquelien Scherpen (Rijksuniversiteit Groningen).
The deadline for submission has been extended to: February 3, 2019. More info in the call for paper: https://www.dcsc.tudelft.nl/~esteur/docs/Call_for_papers_LSS_2019.pdf
The Mini-Symposium “Mathematical Modeling and System-Theoretic Analysis” will be held on February 13, 2019, 10:00 – 17:30 at MF 11-12, Floor 5, Metaforum, Eindhoven University of Technology, The Netherlands
We are happy to announce our three renowned and internationally recognized keynote speakers: Peter Markowich, Volker Mehrmann and Marie-Therese Wolfram. The morning session will be dedicated to continuum modeling for several real life applications and the afternoon session will be dedicated to port-Hamiltonian systems, including modeling, discretization and model reduction aspects. The program also offers sufficient time for networking and discussions.
More details can be accessed at:
http://www.win.tue.nl/casa/meetings/special/mor19/index.html
More details will be available on this web page in early January. Please stay tuned.
Registration:
Participation is free of charge, but registration is mandatory. To register, please fill the google form at: http://goo.gl/wTkeX5. Kindly register by February 4, 2019. For more information, feel free to contact Xingang Cao (x.cao.1@tue.nl) or Harshit Bansal (h.bansal@tue.nl).
The University of Oxford has a position open for a Departmental Lecturer. This is a four year appointment which would be suitable for a post-doc or someone who is close to completing a PhD and who wants to gain experience for a permanent academic position. More details are available at https://www.jobs.ac.uk/job/BOI006/departmental-lecturer-in-control-engineering.
Eindhoven University Technology
Name: Bardia Sharif
Starting date: 01-10-2018
Function: PhD candidate
Email: b.sharif@tue.nl
Group: CST
Supervisors: Dr. ir. Marcel Heertjes & prof.dr.ir.Maurice Heemels
My name is Bardia Sharif and I am 25 years old. I received my BSc and MSc degrees in Electrical engineering at the Eindhoven University of Technology in 2016 and 2018 respectively. For my MSc graduation project I worked on the topic of LPV control of nonlinear systems at the Control Systems group of TUe under the supervision of dr. ir. Roland Tóth. Since October 2018 I have started to work as a PhD candidate in the Control Systems Technology group of the same institution. My current research is concerned with formalization and analysis of Hybrid Integrator Gain Systems (HIGS) for control performance enhancement in linear motion systems. The project is carried out under the supervision of dr.ir.Marcel Heertjes and Prof.dr.ir.Maurice Heemels.
Name: Jordy Senden
Starting date: 01-01-2019
Function: PhD candidate
Email: j.p.f.senden@tue.nl
Group: Control Systems Technology
Supervisor: René van de Molengraft & Herman Bruyninckx
My name is Jordy Senden and I am 27 years old. In 2009 I started my BSc-degree in Mechanical Engineering at Eindhoven University of Technology. After the bachelor, I continued at the TU/e to do a master in Control Systems Technology. For my graduation project I joined the robot-soccer team Tech United. The initial goal of my project was to investigate whether it was possible to give curveballs with the current robots. It soon became clear that the robots would not be able to do so. The project changed into a model-based approach to enable the robots to shoot accurately at a 3D target. After finalising the master thesis my supervisor, René van de Molengraft, suggested that I might be interested in a PDEng Mechatronics Systems Design. This was a great opportunity for me to broaden my knowledge on systems engineering and design. After successfully finishing the two-year PDEng track, I was given the opportunity to do a PhD in the CST-group. During my PhD I will be part of the FlexCRAFT consortium, which will do research on cognitive robots for flexible agro-food technology. The goal is to equip robot technology with generic capabilities in active perception, world modelling, planning and control, and gripping and manipulation to deal with variations in a robust way. I will do this under the supervision of René van de Molengraft and Herman Bruyninckx.
Name: Tomas Meijer
Starting data: 7-1-2019
Function: PhD student
Email: t.j.meijer@tue.nl
Group: Control Systems Technology
Supervisors: Maurice Heemels, Bram de Jager en Victor Dolk
Project: Model predictive control for thermal systems
My name is Tomas Meijer, 24 years old, and I recently started my PhD project in the Control Systems Technology group at Eindhoven University of Technology. I obtained both my BSc-degree and MSc-degree in Electrical Engineering also in Eindhoven. For my Master’s graduation project, I worked on adaptive port-Hamiltonian discretization of distributed parameter systems in the Control Systems group under the supervision of Prof.dr. Siep Weiland and Ir. David van den Hurk. During my PhD I will be working on model predictive control techniques for thermal systems in the Control Systems Technology group and in collaboration with ASML. Besides my interest in technology, I enjoy spending my time travelling, hiking and cycling.
Name: Sjoerd Knippenberg
Starting date: 15-10-2018
Function: PhD candidate
Email: s.c.m.knippenberg@tue.nl
Group: Control Systems Technology
Supervisor: Pascal Etman
Project: Renovation of navigation locks in the Netherlands
My name is Sjoerd Knippenberg, 26 years old and started my PhD project recently at the Control Systems technology group at the Eindhoven University of Technology. After receiving my BSc-degree in Automotive in 2016, I received my MSc-degree in Automotive Technology in September 2018. During my Master’s I have spent three months at the Product Development group of the Technical University of Munich. Here, I worked on the management of complex systems in technical product development using structural metrics, in corporation with Soley GmbH. For my graduation project, I worked on function and activity specification for multi-domain matrix analysis of a Formula Student racecar, under supervision of dr.ir. L.F.P. Etman. During my PhD project, I will investigate the renovation of navigation locks in the Netherlands by applying dependency structure matrix (DSM) techniques. My research is part of the MultiWaterWerk project in corporation with Rijkswaterstaat.
Name: Xiaodong Cheng
Starting date: 01-01-2019
Function: Postdoc
Email: x.cheng@tue.nl
Group: Control Systems
Supervisor: Paul van den Hof
Project: Data-driven modelling in dynamic networks
I received my B.S. and M.S. degrees in system and control engineering from Northwestern Polytechnical University, Xi’an, China, in 2011 and 2014, respectively. From September 2014 to December 2018, I was a Ph.D. student at University of Groningen, The Netherlands, supervised by Prof. Jacquelien M.A. Scherpen. I was working on model reduction of network systems and received my Ph.D. degree in November 2018. Since January 2019, I have been a postdoc researcher with the Department of Electrical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands. My current topic is system identification of dynamic networks.
Delft University of technology
Name: Luka Peternel
Starting date: 15-01-2019
Function: UD
Group: CoR
I was born in 1987 in Ljubljana, Slovenia and grew up in a small mountain village near Idrija in the west of Slovenia. I attended an electrical engineering high school in Ljubljana and continued to study robotics at Faculty of Electrical Engineering of University of Ljubljana. I received a Ph.D. in robotics in 2015. I conducted my Ph.D. studies at Department of Automation, Biocybernetics and Robotics, Jožef Stefan Institute in Ljubljana from 2011 to 2015, and at Department of Brain-Robot Interface, ATR Computational Neuroscience Laboratories in Kyoto, Japan in 2013 and 2014. In 2015 I became a postdoctoral researcher at Human-Robot Interfaces and Physical Interaction Lab, Advanced Robotics, Italian Institute of Technology in Genoa, Italy. Up to this point, my research mainly involved physical human-robot interaction control, exoskeleton control, human motor control and computational neuroscience. In free time I like to travel, do scuba diving, play football, occasionally play guitar and go out with friends for live music (especially jazz).
Name: Yke Bauke Eisma
Starting date: 01-01-2019
Function: PhD candidate
Email: y.b.eisma@tudelft.nl
Group: CoR
Supervisor: dr. ir. J.C.F. de Winter
My name is Yke Bauke Eisma; I am married and have a son of one year old. I live in Rotterdam-south, and travel back and forth by car (soon on an electric scooter though) and have thus plenty of experience with traffic jams. The nice thing is that my PhD (with Dr. de Winter) concerns research on self-driving cars, possibly eliminating the source of my daily irritations with traffic. A bit more detail: my research with Dr. de Winter is about external Human-Machine interfaces for autonomous cars, more specifically: how we should design and evaluate them. But since I have very broad interests in psychological topics, on occasion I will do an off-topic experiment with my trusted eye-tracker, in order to learn how the world around us is real or whether we may live in a simulation. Either way, I hope I will enjoy my time here.
Eindhoven University of Technology
Candidate: A. Morales Medina
Group: Dynamics & Control
Thesis: Cooperative Intersection Control for Autonomous Vehicles
Promotor: Prof.dr. H. Nijmeijer & prof. dr. ir. N. van de Wouw
Date: 10-01-2019
Location: TU/e, auditorium, Senaatzaal
Time: 11 am
University of Groningen
Candidate: M. Dresscher
Group: ENTEG (DPTA)
Thesis: Toward Controlled Ultra-High Vacuum Chemical Vapor Deposition Processes
Promotor: prof. B. Jayawardhana, prof. J.M.A Scherpen en prof. B. J. Kooi
Date: 11-01-2019
Location: Aula Academiegebouw, Broerstraat 5, Groningen
Time: 12:45
Delft University of Technology
Candidate: S. Boersma
Group: DCSC
Thesis: Towards closed-loop dynamical wind farm control: model development and control applications
Promotor: Prof.dr.ir. J.W. van Wingerden
Date: 15-01-2019
Location: Aula Senaatszaal, Mekelweg 5
Time: 10.00
The next issue of this Newsletter will appear in February 2019
We encourage the contributors to provide newsworthy information. In principle, we intend to publish any message offered. However, we reserve the right to edit certain parts of a submission.
Please send your contributions to: m.w.otte@tudelft.nl
You receive this newsletter because you are a member of DISC. If you want to be removed from the mailing list, please let us know by sending an email to m.w.otte@tudelft.nl
