A repository that presents a digital twin model for a robot and how to use it for
failure simulation, reliability modeling, and predictive maintenance.
Bugs and issues
·
Discussions
Below are a few tutorials that help you get started with this repository.
- Digital twins:
- A matlab application for condition-monitoring of a robot
- A digital shadow that synchronizes the digital twin model with the real robot (to be updated)
- Trajectory simulation model in Matlab (to be updated)
- Trajectory planning model in MoveIt (to be updated)
- Reliability and predictive maintenance based on digital twin:
- Training a deep learning model based on digital twin for fault diagnosis (to be updated)
- Using digital twin to detect the temperature drift of the robot motors (see this data challenge)
This project is currently under development. The final objectives are to create open source digital twin model of a robot and codes and tutorials on how to use the digital twin model for failure simulation, reliability modeling, and predictive maintenance.
Below is a list of new features that we are working on right now:
- Add a functionality to simulate the temperature and voltage in the digital twin model.
- Add a use case and tutorial of using the digital twin model to train and fine-tune a machine learning model to predict the failure of the real robot.
The main parts of this repository is organized as follows:
└── robot_digital_twin: This folder contains the digital twin model for the robot.
└── 3d_model_urdf_files: 3D model of the robot in URDF format.
└── catin_ws: A copy of the catkin workspace on the robot. It contains the source codes on the robot side that are related to this repository.
└── condition_monitoring_matlab_ros: A matlab application for condition-monitoring.
└── digital_shadow_python_ros (to be finished): A digital shadow (data flow from the robot to the model) developed in python.
└── trajectory_simulation_matlab (to be finished): A matlab application for simulating trajectories.
└── trajectory_simulation_moveit (to be finished): A moveit application for simulating trajectories.
└── documents: This folder contains the tutorials related to this project.
└── related_projects (to be finished): This folder contains the related projects.
You are welcome to contribute to this project. Please folk the repository, add a your own branch, make your changes on this branche, and submit a pull request. Please describe briefly what you have changed so that we can easily understand and review your changes.
Code and documentation copyright 2023- the authors. Code released under the MIT License.
This work is partially supported by the French Research Council (ANR) JCJC projct Digital Failure Twin under contract number ANR-22-CE10-0004.
This project benifits from the following open source projects:
- To be updated.
Thanks goes to these wonderful people (emoji key):
sonic160 📆 💻 |
ZAPHAKIEL 💻 |
Thomas W 💻 |
KillianMcCourt 💻 |
X1M1C1 💻 |
This project follows the all-contributors specification. Contributions of any kind welcome!