End-to-end robot manipulation using demonstration-guided goal strategies

Cheol Hui Min, Jae Bok Song

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

In deep reinforcement learning, finding the optimal manipulation policy of a multi-DOF manipulator in 3D space requires intricate reward shaping for the agent to find the optimal policy. However, reward shaping requires cumbersome optimization of the reward function based on prior knowledge on robotic tasks to achieve. This makes it desirable to learn various manipulation policies with a simple reward function.In this study, we propose a method that learns the manipulation policy of a manipulator in a sparse reward setting. To this end, Hindsight Experience Replay (HER) is combined with Twin Delayed DDPG (TD3) by applying the goal strategy that incorporates demonstrations for the policy. It is shown that the policy can estimate the joint control command of a 7-DoF manipulator from raw RGB video inputs in sparse reward setting in an end-to-end manner.

Original languageEnglish
Title of host publication2019 16th International Conference on Ubiquitous Robots, UR 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages159-164
Number of pages6
ISBN (Electronic)9781728132327
DOIs
Publication statusPublished - 2019 Jun
Event16th International Conference on Ubiquitous Robots, UR 2019 - Jeju, Korea, Republic of
Duration: 2019 Jun 242019 Jun 27

Publication series

Name2019 16th International Conference on Ubiquitous Robots, UR 2019

Conference

Conference16th International Conference on Ubiquitous Robots, UR 2019
Country/TerritoryKorea, Republic of
CityJeju
Period19/6/2419/6/27

Bibliographical note

Funding Information:
This work was supported by IITP grant funded by the Korea Government MSIT. (No. 2018-0-00622)

Publisher Copyright:
© 2019 IEEE.

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Human-Computer Interaction
  • Mechanical Engineering
  • Control and Optimization

Fingerprint

Dive into the research topics of 'End-to-end robot manipulation using demonstration-guided goal strategies'. Together they form a unique fingerprint.

Cite this