Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning
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 re...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-08-01
|
Series: | Biomimetics |
Subjects: | |
Online Access: | https://www.mdpi.com/2313-7673/8/4/364 |
_version_ | 1797585418565189632 |
---|---|
author | Haoran Sun Linhan Yang Yuping Gu Jia Pan Fang Wan Chaoyang Song |
author_facet | Haoran Sun Linhan Yang Yuping Gu Jia Pan Fang Wan Chaoyang Song |
author_sort | Haoran Sun |
collection | DOAJ |
description | 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. |
first_indexed | 2024-03-11T00:05:50Z |
format | Article |
id | doaj.art-d55fd5e6d6994ded9c3a8e52f4a0c78f |
institution | Directory Open Access Journal |
issn | 2313-7673 |
language | English |
last_indexed | 2024-03-11T00:05:50Z |
publishDate | 2023-08-01 |
publisher | MDPI AG |
record_format | Article |
series | Biomimetics |
spelling | doaj.art-d55fd5e6d6994ded9c3a8e52f4a0c78f2023-11-19T00:22:50ZengMDPI AGBiomimetics2313-76732023-08-018436410.3390/biomimetics8040364Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement LearningHaoran Sun0Linhan Yang1Yuping Gu2Jia Pan3Fang Wan4Chaoyang Song5Department of Mechanical and Energy Engineering, Southern University of Science and Technology, Shenzhen 518055, ChinaDepartment of Mechanical and Energy Engineering, Southern University of Science and Technology, Shenzhen 518055, ChinaDepartment of Mechanical and Energy Engineering, Southern University of Science and Technology, Shenzhen 518055, ChinaDepartment of Computer Science, The University of Hong Kong, Hong Kong SAR, ChinaShenzhen Key Laboratory of Flexible Manufacturing and Robotics, Southern University of Science and Technology, Shenzhen 518055, ChinaDepartment of Mechanical and Energy Engineering, Southern University of Science and Technology, Shenzhen 518055, ChinaLocomotion 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.https://www.mdpi.com/2313-7673/8/4/364loco-manipulationreinforcement learningreconfigurable robot |
spellingShingle | Haoran Sun Linhan Yang Yuping Gu Jia Pan Fang Wan Chaoyang Song Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning Biomimetics loco-manipulation reinforcement learning reconfigurable robot |
title | Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning |
title_full | Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning |
title_fullStr | Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning |
title_full_unstemmed | Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning |
title_short | Bridging Locomotion and Manipulation Using Reconfigurable Robotic Limbs via Reinforcement Learning |
title_sort | bridging locomotion and manipulation using reconfigurable robotic limbs via reinforcement learning |
topic | loco-manipulation reinforcement learning reconfigurable robot |
url | https://www.mdpi.com/2313-7673/8/4/364 |
work_keys_str_mv | AT haoransun bridginglocomotionandmanipulationusingreconfigurableroboticlimbsviareinforcementlearning AT linhanyang bridginglocomotionandmanipulationusingreconfigurableroboticlimbsviareinforcementlearning AT yupinggu bridginglocomotionandmanipulationusingreconfigurableroboticlimbsviareinforcementlearning AT jiapan bridginglocomotionandmanipulationusingreconfigurableroboticlimbsviareinforcementlearning AT fangwan bridginglocomotionandmanipulationusingreconfigurableroboticlimbsviareinforcementlearning AT chaoyangsong bridginglocomotionandmanipulationusingreconfigurableroboticlimbsviareinforcementlearning |