Development of a Digital Twin for the UR3 industrial collaborative robot arm with the Robotiq Hande gripper attachment using ROS: Laying the Foundation for Reinforcement Learning Research and Advanced Academic Exploration
In the context of the Fourth Industrial Revolution, the Ambiente Integrado de Aprendizaje (AIA) lab at the University of the Andes, Colombia, recognized the need to utilize underexploited hardware for research and development. This research aimed to create a digital twin of the UR3 Industrial Collab...
- Autores:
-
Méndez Galvis, Juan Andrés
- Tipo de recurso:
- Trabajo de grado de pregrado
- Fecha de publicación:
- 2023
- Institución:
- Universidad de los Andes
- Repositorio:
- Séneca: repositorio Uniandes
- Idioma:
- eng
- OAI Identifier:
- oai:repositorio.uniandes.edu.co:1992/69293
- Acceso en línea:
- http://hdl.handle.net/1992/69293
- Palabra clave:
- Digital Twin
Internet of Things
UR3 Industrial Collaborative Robot
Reinforcement Learning
ROS
Gazebo
Software Development
Object Recognition
Robotics
Dockerization
Incremental Development
Simulation
Open-source software
Ingeniería
- Rights
- openAccess
- License
- Atribución 4.0 Internacional
| Summary: | In the context of the Fourth Industrial Revolution, the Ambiente Integrado de Aprendizaje (AIA) lab at the University of the Andes, Colombia, recognized the need to utilize underexploited hardware for research and development. This research aimed to create a digital twin of the UR3 Industrial Collaborative Robot, mimicking a pick-and-place task, to make experimentation accessible to researchers, students, and curiosity-driven individuals. Using an incremental development methodology in Robot Operating System (ROS) Gazebo and adhering to best practices in Python, a digital twin with integrated telemetry and object recognition modules was successfully developed. A detailed architecture was also outlined, bridging both real hardware and simulated environments. The project resulted in a digital twin that accurately reflected the real hardware, accessible via Dockerization, and contributed to the AIA lab¿s capabilities. Challenges were encountered in accurately simulating specific behaviors, such as the grasping task, highlighting areas for future research. This work marks a significant step towards the integration of digital twin technology in Reinforcement learning (RL) applications and showcases the potential of digitalization in enhancing experimental accessibility and flexibility in robotics. |
|---|