Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing

Welding robots are essential in modern manufacturing, providing high precision and efficiency in welding processes. To optimize their performance and minimize errors, accurate simulation of their behavior is crucial. This paper presents a novel approach to enhance the simulation of welding robots us...

Full description

Saved in:
Bibliographic Details
Main Authors: Than Le, Le Quang Vinh, Van Huy Pham
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/11/3349
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849722232721375232
author Than Le
Le Quang Vinh
Van Huy Pham
author_facet Than Le
Le Quang Vinh
Van Huy Pham
author_sort Than Le
collection DOAJ
description Welding robots are essential in modern manufacturing, providing high precision and efficiency in welding processes. To optimize their performance and minimize errors, accurate simulation of their behavior is crucial. This paper presents a novel approach to enhance the simulation of welding robots using the Virtual, Augmented, and Mixed Reality (VAM) simulation platform. The VAM platform offers a dynamic and versatile environment that enables a detailed and realistic representation of welding robot actions, interactions, and responses. By integrating VAM with existing simulation techniques, we aim to improve the fidelity and realism of the simulations. Furthermore, to accelerate the learning and optimization of the welding robot’s behavior, we incorporate deep reinforcement learning (DRL) techniques. Specifically, DRL is utilized for task offloading and trajectory planning, allowing the robot to make intelligent decisions in real-time. This integration not only enhances the simulation’s accuracy but also improves the robot’s operational efficiency in smart manufacturing environments. Our approach demonstrates the potential of combining advanced simulation platforms with machine learning to advance the capabilities of industrial robots. In addition, experimental results show that ANFIS achieves higher accuracy and faster convergence compared to traditional control strategies such as PID and FLC.
format Article
id doaj-art-2738eae2c2aa4dd9b3a289236c81f892
institution DOAJ
issn 1424-8220
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-2738eae2c2aa4dd9b3a289236c81f8922025-08-20T03:11:24ZengMDPI AGSensors1424-82202025-05-012511334910.3390/s25113349Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart ManufacturingThan Le0Le Quang Vinh1Van Huy Pham2Institute of Engineering and Technology, Thu Dau Mot University, Thu Dau Mot 75100, VietnamWisdom Research, Ho Chi Minh City 700000, VietnamFaculty of Information Technology, Ton Duc Thang University, Ho Chi Minh City 700000, VietnamWelding robots are essential in modern manufacturing, providing high precision and efficiency in welding processes. To optimize their performance and minimize errors, accurate simulation of their behavior is crucial. This paper presents a novel approach to enhance the simulation of welding robots using the Virtual, Augmented, and Mixed Reality (VAM) simulation platform. The VAM platform offers a dynamic and versatile environment that enables a detailed and realistic representation of welding robot actions, interactions, and responses. By integrating VAM with existing simulation techniques, we aim to improve the fidelity and realism of the simulations. Furthermore, to accelerate the learning and optimization of the welding robot’s behavior, we incorporate deep reinforcement learning (DRL) techniques. Specifically, DRL is utilized for task offloading and trajectory planning, allowing the robot to make intelligent decisions in real-time. This integration not only enhances the simulation’s accuracy but also improves the robot’s operational efficiency in smart manufacturing environments. Our approach demonstrates the potential of combining advanced simulation platforms with machine learning to advance the capabilities of industrial robots. In addition, experimental results show that ANFIS achieves higher accuracy and faster convergence compared to traditional control strategies such as PID and FLC.https://www.mdpi.com/1424-8220/25/11/3349arc welding robotVAMdigital twinsunityARsimulation
spellingShingle Than Le
Le Quang Vinh
Van Huy Pham
Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing
Sensors
arc welding robot
VAM
digital twins
unity
AR
simulation
title Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing
title_full Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing
title_fullStr Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing
title_full_unstemmed Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing
title_short Virtual, Augmented, and Mixed Reality Robotics-Assisted Deep Reinforcement Learning Towards Smart Manufacturing
title_sort virtual augmented and mixed reality robotics assisted deep reinforcement learning towards smart manufacturing
topic arc welding robot
VAM
digital twins
unity
AR
simulation
url https://www.mdpi.com/1424-8220/25/11/3349
work_keys_str_mv AT thanle virtualaugmentedandmixedrealityroboticsassisteddeepreinforcementlearningtowardssmartmanufacturing
AT lequangvinh virtualaugmentedandmixedrealityroboticsassisteddeepreinforcementlearningtowardssmartmanufacturing
AT vanhuypham virtualaugmentedandmixedrealityroboticsassisteddeepreinforcementlearningtowardssmartmanufacturing