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.
@article{arxiv.2112.05621,
title = {Reward-Based Environment States for Robot Manipulation Policy Learning},
author = {Cédérick Mouliets and Isabelle Ferrané and Heriberto Cuayáhuitl},
journal= {arXiv preprint arXiv:2112.05621},
year = {2021}
}
Comments
NeurIPS Workshop on Deployable Decision Making in Embodied Systems, 2021