The code for the project Multi-agent Motion Planning on the Hybrid Graph.
The core functionalities are encapsulated in the panav Python package.
- Python 3.8/3.9/3.10
- Jupyter lab
- matplotlib
- faiss: a library to efficiently compute nearest neighbors(developed by Meta Research), necessary for RRT algorithms. Install the CPU version by
conda install -c pytorch faiss-cpu. Install the GPU version byconda install -c pytorch faiss-gpu. - polytope: a Python package that allows flexible construction of convex and non-convex polytopes, called polytope, which is part of the TuLip control toolbox. Install by
pip install polytope. - cvxopt: a convex optimization library. Install by
conda install -c conda-forge cvxopt. Required by pypoman. - pypoman: this library implements common operations over convex polyhedra such as polytope projection, double description (conversion between halfspace and vertex representations), computing the Chebyshev center, etc. Install by
pip install pypoman. - shapely: a powerful library for geometry shape manipulations. Useful in visualization. Install by
conda install -c conda-forge shapely - cvxpy: a convenient Python interface to write and solve convex and mix-interger programming problems.
pip install cvxpy. - Gurobi: a semi-commericalized optimization solver. Install its python interface + gurobi license. No need to install the entire Gurobi software.
- Install Gurobi Python interface, necessary for using Gurobi as the backend solver in cvxpy: https://www.gurobi.com/documentation/10.0/quickstart_windows/cs_anaconda_and_grb_conda_.html.
- Create a Gurobi account, request for an academic license, run grbt get key to install the license on your computer. https://www.gurobi.com/login/
- Networkx: used to construct the high-level MAMP search tree, as well as determining the order in a partial ordering. Install by
pip install networkx
- Virtual navigation environment construction and visualization.
- Global multi-agent motion planning.
- Local path tracking & distributed multi-agent collision avoidance.
- Experiments written on Python notebooks.