An implementation of the DGPMP2-ND differentiable motion planner.
DGPMP2-ND is a modified and extended variant of Differentiable Gaussian Process Motion Planning DGPMP2 [1]. The code in this repo is based on this implementation by Mustafa Mukadam.
- Install a recent version of Python. This code was tested with Python 3.11.9 under Windows. Create a virtual environment if desired.
- Clone this repository:
git clone [email protected]:benjaminalt/dgpmp2-nd.git - Install this repository into the currently active Python environment:
cd dgpmp2-nd
pip install .This will automatically install all binary dependencies, as well as differentiable-robot-model as an additional dependency.
A usage example and demo can be found under demos/planning_demo.ipynb.