Leveraging large language models for comprehensive locomotion control in humanoid robots design

This paper investigates the utilization of large language models (LLMs) for the comprehensive control of humanoid robot locomotion. Traditional reinforcement learning (RL) approaches for robot locomotion are resource-intensive and rely heavily on manually designed reward functions. To address these...

Full description

Saved in:
Bibliographic Details
Main Authors: Shilong Sun, Chiyao Li, Zida Zhao, Haodong Huang, Wenfu Xu
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Biomimetic Intelligence and Robotics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2667379724000457
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850101045614608384
author Shilong Sun
Chiyao Li
Zida Zhao
Haodong Huang
Wenfu Xu
author_facet Shilong Sun
Chiyao Li
Zida Zhao
Haodong Huang
Wenfu Xu
author_sort Shilong Sun
collection DOAJ
description This paper investigates the utilization of large language models (LLMs) for the comprehensive control of humanoid robot locomotion. Traditional reinforcement learning (RL) approaches for robot locomotion are resource-intensive and rely heavily on manually designed reward functions. To address these challenges, we propose a method that employs LLMs as the primary designer to handle key aspects of locomotion control, such as trajectory planning, inverse kinematics solving, and reward function design. By using user-provided prompts, LLMs generate and optimize code, reducing the need for manual intervention. Our approach was validated through simulations in Unity, demonstrating that LLMs can achieve human-level performance in humanoid robot control. The results indicate that LLMs can simplify and enhance the development of advanced locomotion control systems for humanoid robots.
format Article
id doaj-art-676ecb01bc4e47e480a50783ffd991b6
institution DOAJ
issn 2667-3797
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Biomimetic Intelligence and Robotics
spelling doaj-art-676ecb01bc4e47e480a50783ffd991b62025-08-20T02:40:08ZengElsevierBiomimetic Intelligence and Robotics2667-37972024-12-014410018710.1016/j.birob.2024.100187Leveraging large language models for comprehensive locomotion control in humanoid robots designShilong Sun0Chiyao Li1Zida Zhao2Haodong Huang3Wenfu Xu4School of Mechanical Engineering and Automation, Harbin Institute of Technology, Shenzhen 518055, China; Guangdong Provincial Key Laboratory of Intelligent Morphing Mechanisms and Adaptive Robotics, Shenzhen 518055, China; Corresponding author at:School of Mechanical Engineering and Automation, Harbin Institute of Technology, Shenzhen 518055, ChinaSchool of Mechanical Engineering and Automation, Harbin Institute of Technology, Shenzhen 518055, ChinaSchool of Mechanical Engineering and Automation, Harbin Institute of Technology, Shenzhen 518055, ChinaSchool of Mechanical Engineering and Automation, Harbin Institute of Technology, Shenzhen 518055, China; Guangdong Provincial Key Laboratory of Intelligent Morphing Mechanisms and Adaptive Robotics, Shenzhen 518055, China; Key University Laboratory of Mechanism & Machine Theory and Intelligent Unmanned Systems of Guangdong, Shenzhen 518055, ChinaThis paper investigates the utilization of large language models (LLMs) for the comprehensive control of humanoid robot locomotion. Traditional reinforcement learning (RL) approaches for robot locomotion are resource-intensive and rely heavily on manually designed reward functions. To address these challenges, we propose a method that employs LLMs as the primary designer to handle key aspects of locomotion control, such as trajectory planning, inverse kinematics solving, and reward function design. By using user-provided prompts, LLMs generate and optimize code, reducing the need for manual intervention. Our approach was validated through simulations in Unity, demonstrating that LLMs can achieve human-level performance in humanoid robot control. The results indicate that LLMs can simplify and enhance the development of advanced locomotion control systems for humanoid robots.http://www.sciencedirect.com/science/article/pii/S2667379724000457Humanoid robotsLarge language modelsLocomotion controlReinforcement learning
spellingShingle Shilong Sun
Chiyao Li
Zida Zhao
Haodong Huang
Wenfu Xu
Leveraging large language models for comprehensive locomotion control in humanoid robots design
Biomimetic Intelligence and Robotics
Humanoid robots
Large language models
Locomotion control
Reinforcement learning
title Leveraging large language models for comprehensive locomotion control in humanoid robots design
title_full Leveraging large language models for comprehensive locomotion control in humanoid robots design
title_fullStr Leveraging large language models for comprehensive locomotion control in humanoid robots design
title_full_unstemmed Leveraging large language models for comprehensive locomotion control in humanoid robots design
title_short Leveraging large language models for comprehensive locomotion control in humanoid robots design
title_sort leveraging large language models for comprehensive locomotion control in humanoid robots design
topic Humanoid robots
Large language models
Locomotion control
Reinforcement learning
url http://www.sciencedirect.com/science/article/pii/S2667379724000457
work_keys_str_mv AT shilongsun leveraginglargelanguagemodelsforcomprehensivelocomotioncontrolinhumanoidrobotsdesign
AT chiyaoli leveraginglargelanguagemodelsforcomprehensivelocomotioncontrolinhumanoidrobotsdesign
AT zidazhao leveraginglargelanguagemodelsforcomprehensivelocomotioncontrolinhumanoidrobotsdesign
AT haodonghuang leveraginglargelanguagemodelsforcomprehensivelocomotioncontrolinhumanoidrobotsdesign
AT wenfuxu leveraginglargelanguagemodelsforcomprehensivelocomotioncontrolinhumanoidrobotsdesign