Research on AGV Path Planning Based on Improved DQN Algorithm

Traditional deep reinforcement learning methods suffer from slow convergence speeds and poor adaptability in complex environments and are prone to falling into local optima in AGV system applications. To address these issues, in this paper, an adaptive path planning algorithm with an improved Deep Q...

Full description

Saved in:
Bibliographic Details
Main Authors: Qian Xiao, Tengteng Pan, Kexin Wang, Shuoming Cui
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/15/4685
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849770499153854464
author Qian Xiao
Tengteng Pan
Kexin Wang
Shuoming Cui
author_facet Qian Xiao
Tengteng Pan
Kexin Wang
Shuoming Cui
author_sort Qian Xiao
collection DOAJ
description Traditional deep reinforcement learning methods suffer from slow convergence speeds and poor adaptability in complex environments and are prone to falling into local optima in AGV system applications. To address these issues, in this paper, an adaptive path planning algorithm with an improved Deep Q Network algorithm called the B-PER DQN algorithm is proposed. Firstly, a dynamic temperature adjustment mechanism is constructed, and the temperature parameters in the Boltzmann strategy are adaptively adjusted by analyzing the change trend of the recent reward window. Next, the Priority experience replay mechanism is introduced to improve the training efficiency and task diversity through experience grading sampling and random obstacle configuration. Then, a refined multi-objective reward function is designed, combined with direction guidance, step punishment, and end point reward, to effectively guide the agent in learning an efficient path. Our experimental results show that, compared with other algorithms, the improved algorithm proposed in this paper achieves a higher success rate and faster convergence in the same environment and represents an efficient and adaptive solution for reinforcement learning for path planning in complex environments.
format Article
id doaj-art-3243a3d7e2ab467aa3d85ea491d3c92c
institution DOAJ
issn 1424-8220
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-3243a3d7e2ab467aa3d85ea491d3c92c2025-08-20T03:02:58ZengMDPI AGSensors1424-82202025-07-012515468510.3390/s25154685Research on AGV Path Planning Based on Improved DQN AlgorithmQian Xiao0Tengteng Pan1Kexin Wang2Shuoming Cui3School of Intelligent Science Information Engineering, Shenyang University, Shenyang 110044, ChinaSchool of Intelligent Science Information Engineering, Shenyang University, Shenyang 110044, ChinaSchool of Intelligent Science Information Engineering, Shenyang University, Shenyang 110044, ChinaSchool of Intelligent Science Information Engineering, Shenyang University, Shenyang 110044, ChinaTraditional deep reinforcement learning methods suffer from slow convergence speeds and poor adaptability in complex environments and are prone to falling into local optima in AGV system applications. To address these issues, in this paper, an adaptive path planning algorithm with an improved Deep Q Network algorithm called the B-PER DQN algorithm is proposed. Firstly, a dynamic temperature adjustment mechanism is constructed, and the temperature parameters in the Boltzmann strategy are adaptively adjusted by analyzing the change trend of the recent reward window. Next, the Priority experience replay mechanism is introduced to improve the training efficiency and task diversity through experience grading sampling and random obstacle configuration. Then, a refined multi-objective reward function is designed, combined with direction guidance, step punishment, and end point reward, to effectively guide the agent in learning an efficient path. Our experimental results show that, compared with other algorithms, the improved algorithm proposed in this paper achieves a higher success rate and faster convergence in the same environment and represents an efficient and adaptive solution for reinforcement learning for path planning in complex environments.https://www.mdpi.com/1424-8220/25/15/4685automatic guided vehiclepath planningdeep Q networkdeep reinforcement learning
spellingShingle Qian Xiao
Tengteng Pan
Kexin Wang
Shuoming Cui
Research on AGV Path Planning Based on Improved DQN Algorithm
Sensors
automatic guided vehicle
path planning
deep Q network
deep reinforcement learning
title Research on AGV Path Planning Based on Improved DQN Algorithm
title_full Research on AGV Path Planning Based on Improved DQN Algorithm
title_fullStr Research on AGV Path Planning Based on Improved DQN Algorithm
title_full_unstemmed Research on AGV Path Planning Based on Improved DQN Algorithm
title_short Research on AGV Path Planning Based on Improved DQN Algorithm
title_sort research on agv path planning based on improved dqn algorithm
topic automatic guided vehicle
path planning
deep Q network
deep reinforcement learning
url https://www.mdpi.com/1424-8220/25/15/4685
work_keys_str_mv AT qianxiao researchonagvpathplanningbasedonimproveddqnalgorithm
AT tengtengpan researchonagvpathplanningbasedonimproveddqnalgorithm
AT kexinwang researchonagvpathplanningbasedonimproveddqnalgorithm
AT shuomingcui researchonagvpathplanningbasedonimproveddqnalgorithm