[robotics-worldwide] [meetings] CFP: Tutorial on Dynamical System-based Learning from Demonstration (LfD)

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

[robotics-worldwide] [meetings] CFP: Tutorial on Dynamical System-based Learning from Demonstration (LfD)

Aude Billard
Call for Participation

RSS 2018 Tutorial on Dynamical System-based Learning from Demonstration
June 29, 2018 (9am-12:30pm)
Carnegie Mellon University, Pittsburgh, PA, USA

Website: https://urldefense.proofpoint.com/v2/url?u=https-3A__epfl-2Dlasa.github.io_TutorialRSS2018.io_&d=DwICaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=0w3solp5fswiyWF2RL6rSs8MCeFamFEPafDTOhgTfYI&m=aw0IT289S-TmAHYnVylbPW62PzA9Sim433RCzEh9D7c&s=_AXdj9JiJB6r-05c49wS7UXW7Yy9LT8ln5cDcOYIDBw&e=

This tutorial will introduce students to techniques used to program
robots from human demonstration by means of Dynamical Systems (DS). The
use of DS for motion planning problems in robotics has become popular
thanks to their ability to generate on-line motion plans inherently
robust to changes in dynamic environments. In recent years we have been
focusing on formulating DS to model robotic tasks that can be learned
from demonstrations (LfD). We have used our DS-based learning techniques
in a plethora of robotic applications, from executing simple
point-to-point motions, such as pick-and-place and imitating motion
patterns to more dynamic scenarios, such as generating golf swings,
obstacle avoidance and even catching objects in flight. These techniques
have been further extended to learn more complex tasks of repetitive
nature, from sequential point-to-point motions to peeling vegetables or
rolling pizza dough.

In the first lecture of the tutorial, we will introduce various
techniques used to learn autonomous Dynamical System motion
representations from human demonstrations. The second lecture will focus
on the introduction of modulation strategies to modify a learned
behavior locally. We will particularly focus on (i) local motion
refinement, (ii) obstacle avoidance and (ii) non-contact/contact
transitions. Finally, in the third lecture we will present novel
techniques for impedance/hybrid force control with Dynamical Systems.

Each lecture includes a hands-on practice session via computer-based
simulations in MATLAB. The code for our exercises is compatible with
MATLAB versions R2015a-2017a. Please bring your own laptop with one of
the compatible versions if you wish to participate in these sessions.

If you are interested in attending this tutorial please register at the
following link:

No personal information is required. We solely seek to have an estimate
of the number of attendees for proper room assignment.


- Nadia Figueroa, Learning Algorithms and Systems Laboratory, EPFL
- Seyed Sina Mirrazavi Salehian, Learning Algorithms and Systems
Laboratory, EPFL
- Aude Billard, Learning Algorithms and Systems Laboratory, EPFL
- Klas Kronander, Vicarious AI
- Lukas Huber, Learning Algorithms and Systems Laboratory, EPFL


For any questions or queries contact Nadia Figueroa
([hidden email])



Prof. Aude Billard
LASA laboratory, https://urldefense.proofpoint.com/v2/url?u=http-3A__lasa.epfl.ch&d=DwICaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=0w3solp5fswiyWF2RL6rSs8MCeFamFEPafDTOhgTfYI&m=aw0IT289S-TmAHYnVylbPW62PzA9Sim433RCzEh9D7c&s=ITWpVudtmf-8I8Rq8USIANsgbZQY3jOyTymJZa8L3mE&e=
EPFL - Ecole Polytechnique Federale de Lausanne

Mail to: Station 9, 1015 Lausanne, Switzerland
Email: [hidden email]
Tel: +41-21-693-5464
Fax: +41-21-694-7850

robotics-worldwide mailing list
[hidden email]