Endre søk
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
Deep kernels for optimizing locomotion controllers
KTH, Skolan för datavetenskap och kommunikation (CSC), Robotik, perception och lärande, RPL. (Robotik, perception och lärande, RPL, Robotics, perception and learning, RPL)ORCID-id: 0000-0002-3018-2445
Robotics Institute, School of Computer Science, Carnegie Mellon University, USA.
Robotics Institute, School of Computer Science, Carnegie Mellon University, USA.
2017 (engelsk)Inngår i: Proceedings of the 1st Annual Conference on Robot Learning, PMLR , 2017Konferansepaper, Publicerat paper (Fagfellevurdert)
Abstract [en]

Sample efciency is important when optimizing parameters of locomotion controllers, since hardware experiments are time consuming and expensive. Bayesian Optimization, a sample-efcient optimization framework, has recently been widely applied to address this problem, but further improvements in sample efciency are needed for practical applicability to real-world robots and highdimensional controllers. To address this, prior work has proposed using domain expertise for constructing custom distance metrics for locomotion. In this work we show how to learn such a distance metric automatically. We use a neural network to learn an informed distance metric from data obtained in high-delity simulations. We conduct experiments on two different controllers and robot architectures. First, we demonstrate improvement in sample efciency when optimizing a 5-dimensional controller on the ATRIAS robot hardware. We then conduct simulation experiments to optimize a 16-dimensional controller for a 7-link robot model and obtain signicant improvements even when optimizing in perturbed environments. This demonstrates that our approach is able to enhance sample efciency for two different controllers, hence is a tting candidate for further experiments on hardware in the future. Keywor

sted, utgiver, år, opplag, sider
PMLR , 2017.
Serie
Proceedings of Machine Learning Research
Emneord [en]
Bayesian Optimization, Simulator-to-Robot Transfer, Bipedal Locomotion
HSV kategori
Identifikatorer
URN: urn:nbn:se:kth:diva-220510OAI: oai:DiVA.org:kth-220510DiVA, id: diva2:1168668
Konferanse
1st Annual Conference on Robot Learning (CoRL)
Forskningsfinansiär
Knut and Alice Wallenberg Foundation
Merknad

QC 20180108

Tilgjengelig fra: 2017-12-21 Laget: 2017-12-21 Sist oppdatert: 2018-01-08bibliografisk kontrollert

Open Access i DiVA

fulltext(6908 kB)53 nedlastinger
Filinformasjon
Fil FULLTEXT01.pdfFilstørrelse 6908 kBChecksum SHA-512
b5a1e6cccb4d29792e2b2700243c819b9b9dd8de80426e0f0ed15da10efebb745e9767383ba08052a81ba5b9d4858b303b547f06fe9412ef6adcd992160af145
Type fulltextMimetype application/pdf

Andre lenker

Published version

Søk i DiVA

Av forfatter/redaktør
Antonova, Rika
Av organisasjonen

Søk utenfor DiVA

GoogleGoogle Scholar
Totalt: 53 nedlastinger
Antall nedlastinger er summen av alle nedlastinger av alle fulltekster. Det kan for eksempel være tidligere versjoner som er ikke lenger tilgjengelige

urn-nbn

Altmetric

urn-nbn
Totalt: 385 treff
RefereraExporteraLink to record
Permanent link

Direct link
Referera
Referensformat
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annet format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annet språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf