Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning


Haoran Sun, Linhan Yang, Yuping Gu, Jia Pan, Fang Wan, Chaoyang Song: Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning. In: Biomimetics, vol. 8, no. 4, pp. 364, 2023.

Abstract

Locomotion and manipulation are two essential skills in robotics but are often divided or decoupled into two separate problems. It is widely accepted that the topological duality between multi-legged locomotion and multi-fingered manipulation shares an intrinsic model. However, a lack of research remains to identify the data-driven evidence for further research. This paper explores a unified formulation of the loco-manipulation problem using reinforcement learning (RL) by reconfiguring robotic limbs with an overconstrained design into multi-legged and multi-fingered robots. Such design reconfiguration allows for adopting a co-training architecture for reinforcement learning towards a unified loco-manipulation policy. As a result, we find data-driven evidence to support the transferability between locomotion and manipulation skills using a single RL policy with a multilayer perceptron or graph neural network. We also demonstrate the Sim2Real transfer of the learned loco-manipulation skills in a robotic prototype. This work expands the knowledge frontiers on loco-manipulation transferability with learning-based evidence applied in a novel platform with overconstrained robotic limbs.

BibTeX (Download)

@article{Sun2023BridgingLocomotion,
title = {Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning},
author = {Haoran Sun and Linhan Yang and Yuping Gu and Jia Pan and Fang Wan and Chaoyang Song},
doi = {10.3390/biomimetics8040364},
year  = {2023},
date = {2023-08-14},
urldate = {2023-08-14},
journal = {Biomimetics},
volume = {8},
number = {4},
pages = {364},
abstract = {Locomotion and manipulation are two essential skills in robotics but are often divided or decoupled into two separate problems. It is widely accepted that the topological duality between multi-legged locomotion and multi-fingered manipulation shares an intrinsic model. However, a lack of research remains to identify the data-driven evidence for further research. This paper explores a unified formulation of the loco-manipulation problem using reinforcement learning (RL) by reconfiguring robotic limbs with an overconstrained design into multi-legged and multi-fingered robots. Such design reconfiguration allows for adopting a co-training architecture for reinforcement learning towards a unified loco-manipulation policy. As a result, we find data-driven evidence to support the transferability between locomotion and manipulation skills using a single RL policy with a multilayer perceptron or graph neural network. We also demonstrate the Sim2Real transfer of the learned loco-manipulation skills in a robotic prototype. This work expands the knowledge frontiers on loco-manipulation transferability with learning-based evidence applied in a novel platform with overconstrained robotic limbs.},
keywords = {Biomimetics, Corresponding Author, JCR Q1},
pubstate = {published},
tppubtype = {article}
}