HJARL (Hamilton-Jacobi-guided Adversaril RL)

Code of the paper Learning Robust Policies via Interpretable Hamilton-Jacobi Reachability-Guided Disturbances that is accepted by ICRA, 2025.

This code is based on the repo safe-control-gym.

To reproduce the paper results, please refer WorkingLogs. Detailed instructions of training will come soon.

@article{hu2024learning,
  title={Learning robust policies via interpretable hamilton-jacobi reachability-guided disturbances},
  author={Hu, Hanyang and Zhang, Xilun and Lyu, Xubo and Chen, Mo},
  journal={arXiv preprint arXiv:2409.19746},
  year={2024}
}