Session: CIE–31: Graduate Student Poster Symposium

Paper Number: 147983

147983 - A Study on Digital Human Modeling for Metaverse Factory 

Digital twins are increasingly becoming integral in various industrial fields, playing a pivotal role in the real-time monitoring, simulation, and control of physical entities. Their implementation in industrial settings, characterized by a plethora of equipment, sensors, and complex industrial environments, has drawn significant attention due to the vast amount of data these elements generate. Typically, digital twin applications are centered around entities that inherently produce digital data. However, a notable gap exists in digital twin frameworks for entities that do not directly generate digital data, such as human workers. Addressing this gap, our research leverages mixed reality technology to seamlessly integrate both digital and non-digital data-generating entities within a manufacturing context. This study introduces a novel approach that involves monitoring the movements and postures of workers in real-time using digital twins and effectively transmitting this information to a virtual environment. By utilizing advanced AI-based posture estimation techniques alongside YOLOv5 for real-time spatial coordinate estimation of detected objects, this research has culminated in the creation of a metaverse factory. This virtual factory setting enables comprehensive monitoring and analysis of the manufacturing process, significantly enhancing production efficiency and facilitating informed decision-making. Expected to have broad applications, this technology promises substantial improvements not only in the manufacturing sector but across various industries, driving forward the integration of physical and virtual worlds and setting new standards for industrial operations.

Presenting Author: One Chang Kim Sungkyunkwan Univ

Presenting Author Biography: I received my bachelor's degree from Sungkyunkwan University, majoring in System Management Engineering with a dual major in Mechanical Engineering. Implementing Class Activation Map (CAM), my undergraduate thesis was "Development of a Robot Spot Welding Process Defect Diagnosis System Using Vision Sensors and Deep Learning" which sparked my interest in vision systems. This led me to commence my master's studies in Intelligent Robotics at Sungkyunkwan University. My current research interests lie in vision-based technologies, and I am actively working on projects such as the development of a vision inspection system for knitted fabrics and real-time environmental status sharing based on vision. I plan to continue researching application methods and utilizations of vision technology.

Authors:

One Chang Kim Sungkyunkwan Univ
Sang Won Lee Sungkyunkwan Univ