Skip to Main content Skip to Navigation
Conference papers

Reward-Based Environment States for Robot Manipulation Policy Learning

Abstract : Training robot manipulation policies is a challenging and open problem in robotics and artificial intelligence. In this paper we propose a novel and compact state representation based on the rewards predicted from an image-based task success classifier. Our experiments—using the Pepper robot in simulation with two deep reinforcement learning algorithms on a grab-and-lift task—reveal that our proposed state representation can achieve up to 97% task success using our best policies.
Document type :
Conference papers
Complete list of metadata
Contributor : Isabelle FERRANÉ Connect in order to contact the contributor
Submitted on : Tuesday, April 5, 2022 - 6:41:42 PM
Last modification on : Monday, July 4, 2022 - 10:17:39 AM
Long-term archiving on: : Wednesday, July 6, 2022 - 7:23:28 PM


Files produced by the author(s)


  • HAL Id : hal-03631670, version 1


Cédérick Mouliets, Isabelle Ferrané, Heriberto Cuayáhuitl. Reward-Based Environment States for Robot Manipulation Policy Learning. NeurIPS Workshop on Deployable Decision Making in Embodied Systems (DDM 2021), University of Toronto, Dec 2021, Toronto, Canada. pp.1-6. ⟨hal-03631670⟩



Record views


Files downloads