Learning deep movement primitives using convolutional neural networks

Affan Pervez, Yuecheng Mao, Dongheui Lee

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

42 Scopus citations

Abstract

Dynamic Movement Primitives (DMPs) are widely used for encoding motion data. Task parameterized DMP (TP-DMP) can adapt a learned skill to different situations. Mostly a customized vision system is used to extract task specific variables. This limits the use of such systems to real world scenarios. This paper proposes a method for combining the DMP with a Convolutional Neural Network (CNN). Our approach preserves the generalization properties associated with a DMP, while the CNN learns the task specific features from the camera images. This eliminates the need to extract the task parameters, by directly utilizing the camera image during the motion reproduction. The performance of the developed approach is demonstrated through a trash cleaning task, executed with a real robot. We also show that by using the data augmentation, the learned sweeping skill can be generalized for arbitrary objects. The experiments show the robustness of our approach for several different settings.

Original languageEnglish
Title of host publication2017 IEEE-RAS 17th International Conference on Humanoid Robotics, Humanoids 2017
PublisherIEEE Computer Society
Pages191-197
Number of pages7
ISBN (Electronic)9781538646786
DOIs
StatePublished - 22 Dec 2017
Externally publishedYes
Event17th IEEE-RAS International Conference on Humanoid Robotics, Humanoids 2017 - Birmingham, United Kingdom
Duration: 15 Nov 201717 Nov 2017

Publication series

NameIEEE-RAS International Conference on Humanoid Robots
ISSN (Print)2164-0572
ISSN (Electronic)2164-0580

Conference

Conference17th IEEE-RAS International Conference on Humanoid Robotics, Humanoids 2017
Country/TerritoryUnited Kingdom
CityBirmingham
Period15/11/1717/11/17

Fingerprint

Dive into the research topics of 'Learning deep movement primitives using convolutional neural networks'. Together they form a unique fingerprint.

Cite this