search

jsll / IROS2018-Hallucinating-Robots

IROS 2018 paper "Hallucinating Robots: Inferring Obstacle Distances from Partial Laser Measurements".
MIT License
3 stars 2 forks source link

readme

Repo

This repository contains code corresponding to:

J. Lundell, F. Verdoja and V. Kyrki, "Hallucinating Robots: Inferring Obstacle Distances from Partial Laser Measurements," 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Madrid, 2018, pp. 4781-4787.

doi: 10.1109/IROS.2018.8594399

preprint: arxiv

Please cite as:

@inproceedings{lundell2018hallucinating,
    author={J. {Lundell} and F. {Verdoja} and V. {Kyrki}},
    booktitle={2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
    title={Hallucinating Robots: Inferring Obstacle Distances from Partial Laser Measurements},
    year={2018},
    pages={4781-4787},
    doi={10.1109/IROS.2018.8594399},
    ISSN={2153-0866}
}

Dependencies

Should be manually installed before setup:

Optional

Setup

  1. Install the dependencies mentioned above.
  2. Clone this repository 
    git clone git@version.aalto.fi:lundelj2/IROS2018-Hallucinating-Robots.git 
cd IROS2018-Hallucinating-Robots/
  3. Fetch the training and test set either by:
    1. Running the installation script install.sh or
    2. Create the folder datasets and manually place the dataset there.
  4. Run a training instance either by:
    1. Running the python script 
      python hallucinating.py
    2. or, alternatively, if the user installed jupyter run the jupyter notebook hallucinating.ipynb