Opendata, web and dolomites

SKILLS4ROBOTS SIGNED

Policy Learning of Motor Skills for Humanoid Robots

Total Cost €

0

EC-Contrib. €

0

Partnership

0

Views

0

Project "SKILLS4ROBOTS" data sheet

The following table provides information about the project.

Coordinator
TECHNISCHE UNIVERSITAT DARMSTADT 

Organization address
address: KAROLINENPLATZ 5
city: DARMSTADT
postcode: 64289
website: www.tu-darmstadt.de

contact info
title: n.a.
name: n.a.
surname: n.a.
function: n.a.
email: n.a.
telephone: n.a.
fax: n.a.

 Coordinator Country Germany [DE]
 Total cost 1˙405˙572 €
 EC max contribution 1˙405˙572 € (100%)
 Programme 1. H2020-EU.1.1. (EXCELLENT SCIENCE - European Research Council (ERC))
 Code Call ERC-2014-STG
 Funding Scheme ERC-STG
 Starting year 2015
 Duration (year-month-day) from 2015-07-01   to  2021-06-30

 Partnership

Take a look of project's partnership.

# participants  country  role  EC contrib. [€] 
1    TECHNISCHE UNIVERSITAT DARMSTADT DE (DARMSTADT) coordinator 1˙405˙572.00

Map

 Project objective

The goal of SKILLS4ROBOTS is to develop a autonomous skill learning system that enables humanoid robots to acquire and improve a rich set of motor skills. This robot skill learning system will allow scaling of motor abilities up to fully anthropomorphic robots while overcoming the current limitations of skill learning systems to only few degrees of freedom. To achieve this goal, it will decompose complex motor skills into simpler elemental movements – called movement primitives – that serve as building blocks for the higher-level movement strategy and the resulting architecture will be able to address arbitrary, highly complex tasks – up to robot table tennis for a humanoid robot. Learned primitives will be superimposed, sequenced and blended.

Four recent breakthroughs in the PI’s research will make this project possible due to successes on the representation of the parametric probabilistic representations of the elementary movements, on probabilistic imitation learning, on relative entropy policy search-based reinforcement learning and on the modular organization of the representation. These breakthroughs will allow create a general, autonomous skill learning system that can learn many different skills in the exact same framework without changing a single line of programmed code. To accomplish this goal, our skill learning system will autonomously extract the necessary movement primitives out of observed trajectories, learn to generalize these primitives to different situations and select, sequence or combine them such that complex behavior can be synthesized out of the primitive building blocks. We will evaluate our autonomous learning framework on a real humanoid robot platform with 60 degrees of freedom and show that it can learn a large variety of new skills.

 Publications

year authors and title journal last update
List of publications.
2017 Belousov, B.; Neumann, G.; Rothkopf, C.A.; Peters, J.
Catching heuristics are optimal control policies
published pages: 1-15, ISSN: , DOI:
Proceedings of the Karniel Thirteenth Computational Motor Control Workshop 2020-02-12
2016 Belousov, B.; Neumann, G.; Rothkopf, C.; Peters, J.
Catching heuristics are optimal control policies
published pages: , ISSN: , DOI:
Advances in Neural Information Processing Systems yearly 2020-02-12
2017 Abdulsamad, H.; Arenz, O.; Peters, J.; Neumann, G.
State-Regularized Policy Search for Linearized Dynamical Systems
published pages: , ISSN: , DOI:
Proceedings of the International Conference on Automated Planning and Scheduling 2020-02-12
2017 Rudolf Lioutikov, Gerhard Neumann, Guilherme Maeda, Jan Peters
Learning movement primitive libraries through probabilistic segmentation
published pages: 879-894, ISSN: 0278-3649, DOI: 10.1177/0278364917713116
The International Journal of Robotics Research 36/8 2020-02-12
2018 Belousov, Boris; Peters, Jan
f-Divergence constrained policy improvement
published pages: , ISSN: , DOI:
2 2020-02-12

Are you the coordinator (or a participant) of this project? Plaese send me more information about the "SKILLS4ROBOTS" project.

For instance: the website url (it has not provided by EU-opendata yet), the logo, a more detailed description of the project (in plain text as a rtf file or a word file), some pictures (as picture files, not embedded into any word file), twitter account, linkedin page, etc.

Send me an  email (fabio@fabiodisconzi.com) and I put them in your project's page as son as possible.

Thanks. And then put a link of this page into your project's website.

The information about "SKILLS4ROBOTS" are provided by the European Opendata Portal: CORDIS opendata.

More projects from the same programme (H2020-EU.1.1.)

DOUBLE-TROUBLE (2020)

Replaying the ‘genome duplication’ tape of life: the importance of polyploidy for adaptation in a changing environment

Read More  

inhibiTOR (2020)

Novel selective mTORC1 inhibitors

Read More  

AllergenDetect (2019)

Comprehensive allergen detection using synthetic DNA libraries

Read More