ARiADNE

Public code and model of ARiADNE: A Reinforcement learning approach using Attention-based Deep Networks for Exploration, which is accepted for the oral presentation at ICRA 2023.

Note: This is a new implementation of ARiADNE. You can find our original implementation in the main branch. We reimplement ARiADNE to optimize the computing time, RAM/VRAM usage, and compatibility with ROS. The trained model can be directly tested in our ARiADNE ROS planner.

Run

Dependencies

We recommend to use conda for package management. Our planner is coded in Python and based on Pytorch. Other than Pytorch, please install following packages by:

pip install scikit-image matplotlib ray tensorboard

We tested our planner in various version of these packages so you can just install the latest one.

Training

Download this repo and go into the folder:

git clone https://github.com/marmotlab/ARiADNE.git
cd ARiADNE

Launch your conda environment if any and run:

python driver.py

The default training code requires around 8GB VRAM and 20G RAM. You can modify the hyperparameters in parameter.py.

Files

• parameters.py Training parameters.
• driver.py Driver of training program, maintain & update the global network.
• runner.py Wrapper of the workers.
• worker.py Interact with environment and collect episode experience.
• model.py Define attention-based network.
• env.py Autonomous exploration environment.
• node_manager.py Manage and update the informative graph.
• quads Quad tree for node indexing provided by Daniel Lindsley.
• sensor.py Simulate the sensor model of Lidar.
• /maps Maps of training environments provided by Chen et al..

Demo of ARiADNE

Cite

If you find our work helpful or enlightening, feel free to cite our paper:

@INPROCEEDINGS{cao2023ariadne,
  author={Cao, Yuhong and Hou, Tianxiang and Wang, Yizhuo and Yi, Xian and Sartoretti, Guillaume},
  booktitle={2023 IEEE International Conference on Robotics and Automation (ICRA)}, 
  title={ARiADNE: A Reinforcement learning approach using Attention-based Deep Networks for Exploration}, 
  year={2023},
  pages={10219-10225},
  doi={10.1109/ICRA48891.2023.10160565}}

Authors

Yuhong Cao
Tianxiang Hou
Yizhuo Wang
Xian Yi
Guillaume Sartoretti