Python toolbox to ease the use of MPC for tracking a goal with the presence of obstacles. Goal and obstacles may be moving according to known dynamics. With the use of the do_mpc library, both linear and nonlinear MPC is possible.
Currently (this repo is still in its inception), only MPC_2DTracker is implemented with an integrated simulator. To run it and look at some examples and its implementation, see src/Tracker_2D/MPC_2DTracker.py. In future releases, the simulator will be separate and the MPC_2DTracker can be used in connection with other (real) systems.
Be sure to install poetry to your system using (for linux at least)
curl -sSL https://install.python-poetry.org | python3 -
Then run poetry install in the root repo directory. This creates a new virtual environment (or uses a currently running one) and installs the packages defined in requirements.txt to it. To access the virtual environment, running poetry shell may be necessary.
Should you require importing new packages, install them to the virtual environment and add them to the package using:
poetry shell
pip install <package_name>
(pip freeze | grep -v "MPC-Tracker")> requirements.txt
poetry add $(cat requirements.txt)
poetry install
- Upgrade setup.py to pyproject.toml
- Simulator separately
- 3D
- Test 2D with more complex dynamics (e.g. differential drive bot incl. non-holonomic constraints)
- (Robust MPC)