STLGame

Website, Preprint.

Install

The implementation has been tested with Python 3.10 under Mac M3. We recommend installing the simulation inside a virtualenv. You can install the environment by running:

git clone git@github.com:shuoyang2000/STLgame.git
cd STLgame
python3 -m venv STLGame_env
source STLGame_env/bin/activate
pip install -r requirements.txt

We use stlcg to compute the gradient of STL formula.

Reproduce the results

Python path

We run all experiments from the project directory so please add the project directory to the PYTHONPATH environment variable:

export PYTHONPATH=$PYTHONPATH:$

Training

The trained model is saved in the folder models/, and trajectory and exploitability data is saved in data/. To change the directories used, change scripts/config.py and scripts/config_3d.py for kinematic and rotor dynamics correspondingly. To reproduce the training, please run (for instance, for autonomous drones case):

python3 scripts/main.py --dynamic rotor

It may take around 10 minutes for each FSP iteration, depending on your machines.

Test Nash policy and best response against seen and unseen opponents

python3 scripts/test_nash.py --dynamic rotor

RL Training for Best Response

python3 scripts/main_rlbr.py --dynamic rotor

Citation and Contact

If you find this work useful, please consider citing:

@article{yang2024STLGame,
  title={STLGame: Signal Temporal Logic Games in Adversarial Multi-Agent Systems},
  author={Yang, Shuo and Zheng, Hongrui and Vasile, Cristian-Ioan and Pappas, George J and Mangharam, Rahul}
  journal={arXiv preprint arXiv:2412.01656},
  year={2024}
}

If you have any question on this repo, please feel free to contact the author Shuo Yang (yangs1 at seas dot upenn dot edu) or raise an issue.