PyCRAM is a plan executive framework for cognitive robotics that enables robust execution of high-level robot plans in partially observable environments. It provides modular, extensible tools for designing, implementing, and executing robot plans, facilitating integration of new functionalities and heterogeneous robot platforms.
The recommended installation method is via pip:
pip install pycram-roboticsFor an alternative installation from source, use the automated script:
curl -s https://raw.githubusercontent.com/cram2/pycram/dev/scripts/install.sh | bashDetailed installation instructions and manual setup guides are available here.
Test PyCRAM directly in your browser via our Virtual Research Building.
PyCRAM supports executing identical high-level plans on different robot platforms. Below is a demonstration of the same plan running on the PR2 and IAI's Boxy:
| Boxy | PR2 |
|---|---|
 |
 |
The full documentation is maintained at Read the Docs.
Source documentation is located in the doc directory. Instructions for building and viewing the documentation can be found in the corresponding README file.
Comprehensive examples are provided as Jupyter Notebooks in the examples folder and documented in the Examples section. Refer to the examples folder's README for instructions on executing these notebooks.
Explore a variety of labs and demonstrations showcasing PyCRAM's capabilities on the Labs page of our virtual building.
To create a custom lab in the virtual building, consult the vrb branch of this repository, which includes detailed setup instructions and templates.
