Digital twin architecture of autonomous industrial robots with joint deviation compensation toward realization of a smart factory

Abstract

Modern manufacturing is transitioning toward autonomous, unmanned, and flexible production environments, where industrial robots (IRs) must operate intelligently while remaining synchronized with other manufacturing assets. This article presents a comprehensive cloud-enabled Digital Twin (DT) architecture for IRs that enables remote task allocation, autonomous ROS-based execution, near real-time monitoring, and synchronized operation with physical manufacturing assets through a unified Cloud–Edge framework. A Unity-based dashboard, integrated with AWS cloud services and edge devices, allows operators to remotely assign manufacturing tasks with varying operation sequences, enabling flexible coordination among IRs, CNC machines, and 3D printers for smart and dark factory applications. To ensure sustained operational accuracy, the proposed architecture incorporates a Kalman filter-based joint deviation compensation framework that estimates and compensates joint deviations caused by wear, backlash, and material degradation. The proposed framework is experimentally validated using two IRs. Experimental results demonstrate autonomous task execution with RMSE values below the encoder resolution and residual deviations within ±0.2° after compensation. Furthermore, a GUM-based measurement uncertainty analysis validates the metrological reliability of the proposed framework, yielding expanded uncertainties between ±0.0362° and ±0.0460°. Overall, the proposed architecture provides a scalable foundation for autonomous, synchronized, and intelligent smart manufacturing systems.