Simulation Based Behavior for a Welfare Robot

Stefan-Daniel Suvei, Leon Bodenhagen, Thomas Nicky Thulesen, Norbert Krüger

Publikation: Kapitel i bog/rapport/konference-proceedingKonferencebidrag i proceedingsForskningpeer review

121 Downloads (Pure)

Abstract

Learning the appropriate parameters which lead to a successful action, even under the effect of noise and inaccuracies stemming from faulty perception or poor sensor calibration is of great importance in service robotics, where the flexibility of the robotic systems is a main requirement. This paper proposes a method inspired by available industrial robotics solutions which makes use of a dynamic simulator and Kernel Density Estimation to find the parameter sets that lead to a successful Peg-in-Hole action. The obtained solution is successfully tested on a real service robot, where the action is used in the context of picking-up and inserting bottles in a crate for depositing.
OriginalsprogEngelsk
TitelSimulation and Modeling Methodologies, Technologies and Applications : 7th International Conference, SIMULTECH 2017 Madrid, Spain, July 26–28, 2017 Revised Selected Papers
RedaktørerMohammad S. Obaidat, Tuncer Ören, Floriano De Rango
ForlagSpringer
Publikationsdato21. nov. 2018
Sider319-334
ISBN (Trykt)9783030014698
ISBN (Elektronisk)9783030014704
DOI
StatusUdgivet - 21. nov. 2018
Begivenhed7th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Madrid, Spanien
Varighed: 26. jul. 201728. jul. 2017
Konferencens nummer: 7
http://www.simultech.org/?y=2017

Konference

Konference7th International Conference on Simulation and Modeling Methodologies, Technologies and Applications
Nummer7
Land/OmrådeSpanien
ByMadrid
Periode26/07/201728/07/2017
Internetadresse
NavnAdvances in Intelligent Systems and Computing
ISSN2194-5357

Fingeraftryk

Dyk ned i forskningsemnerne om 'Simulation Based Behavior for a Welfare Robot'. Sammen danner de et unikt fingeraftryk.

Citationsformater