디지털 트윈 시뮬레이션의 표준화를 위한 계층형 정보 모델 프레임워크

Abstract

Digital twin technology enables high-fidelity virtual replicas of physical robots and environments for simulation and analysis. We propose a layered information model, Control, Robot, and World, to standardize digital twin development for wheeled mobile robots. The model specifies interoperable artifacts (YAML, URDF, USD, SDF) and supports an automated simulation setup in both Isaac Sim and Gazebo. Integrated with ROS 2 and the Nav2 stack, it provides consistent interfaces for control, localization, and path planning. Experiments in a digital twin testbed show stable autonomous navigation in both simulators, demonstrating that the proposed model enables reproducible and consistent operation across heterogeneous platforms.