Skip to content

Imitation learning based vehicle following policies with safety cages

License

Notifications You must be signed in to change notification settings

sampo-kuutti/dnn-driven-av-safety-cages

Repository files navigation

Safe deep neural network-driven autonomous vehicles using software safety cages

This is the repo for paper Safe deep neural network-driven autonomous vehicles using software safety cages. Train an imitation learning policy to control an autonomous agent for vehicle following. Additionally, safety cages can be used to ensure safety and provide additional training signal when the policy makes a mistake.

Installation

Clone the repo

git clone https://github.com/sampo-kuutti/dnn-driven-av-safety-cages

install requirements:

pip install -r requirements.txt

Training the policy

To run the the training run train_sl.py.

Citing the Repo

If you find the code useful in your research or wish to cite it, please use the following BibTeX entry.

@inproceedings{kuutti2019safe,
  title={Safe deep neural network-driven autonomous vehicles using software safety cages},
  author={Kuutti, Sampo and Bowden, Richard and Joshi, Harita and de Temple, Robert and Fallah, Saber},
  booktitle={International Conference on Intelligent Data Engineering and Automated Learning},
  pages={150--160},
  year={2019},
  organization={Springer, Cham}
}

Releases

No releases published

Packages

No packages published

Languages