Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer

Research on trajectory generation algorithms for unmanned ground vehicles (UGVs) has been actively conducted due to the rapid increase in their use across various fields. Trajectory generation for UGVs requires a high level of precision, as various parameters determine the trajectory’s ef...

Full description

Saved in:
Bibliographic Details
Main Authors: Chang-Hun Ji, Gyeonghun Lim, Youn-Hee Han, Sungtae Moon
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10705302/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846159833769705472
author Chang-Hun Ji
Gyeonghun Lim
Youn-Hee Han
Sungtae Moon
author_facet Chang-Hun Ji
Gyeonghun Lim
Youn-Hee Han
Sungtae Moon
author_sort Chang-Hun Ji
collection DOAJ
description Research on trajectory generation algorithms for unmanned ground vehicles (UGVs) has been actively conducted due to the rapid increase in their use across various fields. Trajectory generation for UGVs requires a high level of precision, as various parameters determine the trajectory’s efficiency and safety. Notably, maximum velocity and acceleration are critical factors impacting trajectory performance. In this paper, we propose a novel algorithm that utilizes reinforcement learning to determine the optimal maximum velocity and acceleration in real-time within dynamic environments. The proposed algorithm overcomes the limitations of traditional fixed parameter settings by determining parameters through real-time environmental adaptation. Furthermore, we also propose a PX4-ROS2 based reinforcement learning framework for achieving stable zero-shot sim-to-real transfer. Experimental results in simulation and real-world environments show that the proposed method significantly improves trajectory safety and efficiency while also demonstrating excellent adaptability to changing environments. Furthermore, validation through identical experimental results in both simulation and real-world environments confirms a stable zero-shot sim-to-real transfer.
format Article
id doaj-art-a19ecba8a66f4ce4a730d06bb9ba5d60
institution Kabale University
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-a19ecba8a66f4ce4a730d06bb9ba5d602024-11-23T00:01:45ZengIEEEIEEE Access2169-35362024-01-011217166217167410.1109/ACCESS.2024.347393510705302Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real TransferChang-Hun Ji0https://orcid.org/0009-0003-4372-8675Gyeonghun Lim1Youn-Hee Han2https://orcid.org/0000-0002-5835-7972Sungtae Moon3Department of Future Convergence Engineering, Korea University of Technology and Education, Cheonan-si, South KoreaDepartment of Intelligent System and Robotics, Chungbuk National University, Cheongju-si, South KoreaDepartment of Future Convergence Engineering, Korea University of Technology and Education, Cheonan-si, South KoreaDepartment of Intelligent System and Robotics, Chungbuk National University, Cheongju-si, South KoreaResearch on trajectory generation algorithms for unmanned ground vehicles (UGVs) has been actively conducted due to the rapid increase in their use across various fields. Trajectory generation for UGVs requires a high level of precision, as various parameters determine the trajectory’s efficiency and safety. Notably, maximum velocity and acceleration are critical factors impacting trajectory performance. In this paper, we propose a novel algorithm that utilizes reinforcement learning to determine the optimal maximum velocity and acceleration in real-time within dynamic environments. The proposed algorithm overcomes the limitations of traditional fixed parameter settings by determining parameters through real-time environmental adaptation. Furthermore, we also propose a PX4-ROS2 based reinforcement learning framework for achieving stable zero-shot sim-to-real transfer. Experimental results in simulation and real-world environments show that the proposed method significantly improves trajectory safety and efficiency while also demonstrating excellent adaptability to changing environments. Furthermore, validation through identical experimental results in both simulation and real-world environments confirms a stable zero-shot sim-to-real transfer.https://ieeexplore.ieee.org/document/10705302/Trajectory generationreinforcement learningreal-time parameter optimizationsim-to-real
spellingShingle Chang-Hun Ji
Gyeonghun Lim
Youn-Hee Han
Sungtae Moon
Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer
IEEE Access
Trajectory generation
reinforcement learning
real-time parameter optimization
sim-to-real
title Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer
title_full Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer
title_fullStr Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer
title_full_unstemmed Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer
title_short Real-Time Parameter Control for Trajectory Generation Using Reinforcement Learning With Zero-Shot Sim-to-Real Transfer
title_sort real time parameter control for trajectory generation using reinforcement learning with zero shot sim to real transfer
topic Trajectory generation
reinforcement learning
real-time parameter optimization
sim-to-real
url https://ieeexplore.ieee.org/document/10705302/
work_keys_str_mv AT changhunji realtimeparametercontrolfortrajectorygenerationusingreinforcementlearningwithzeroshotsimtorealtransfer
AT gyeonghunlim realtimeparametercontrolfortrajectorygenerationusingreinforcementlearningwithzeroshotsimtorealtransfer
AT younheehan realtimeparametercontrolfortrajectorygenerationusingreinforcementlearningwithzeroshotsimtorealtransfer
AT sungtaemoon realtimeparametercontrolfortrajectorygenerationusingreinforcementlearningwithzeroshotsimtorealtransfer