Industrial Digital Twin Integration with ERP and Computer Vision for Real-Time Synchronization

Abstract

This paper presents a practical realization of a dual-layered digital twin framework that integrates both an Enterprise Resource Planning (ERP) system and a vision-based perception system to form a unified simulation environment. The project is structured around two main contributions: first, the integration of a customized Odoo-based ERP system with a dynamic AnyLogic simulation model; and second, the replacement of physical sensors with a computer vision pipeline built using YOLOv8, enabling real-time object tracking and event detection along a production line. The system was deployed and tested in a controlled laboratory environment featuring a six-station manufacturing cell. Data exchange was automated using file-based triggers between systems, ensuring continuous synchronization between the ERP database, physical production line, and simulation twin. A conceptual framework is also introduced, illustrating how the system supports all four levels of industrial analytics: descriptive, diagnostic, predictive, and prescriptive. This work demonstrates the feasibility and future potential of unifying enterprise intelligence with AI-based perception in a modular, scalable digital twin architecture tailored for Industry 4.0 environments.