weakipedian / LogGPisMap

Faithful Euclidean distance field from Log-Gaussian Process Implicit Surfaces

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

LogGPisMap

This repository contains source code for LogGPisMap, introduced in the paper Faithful Euclidean distance field from Log-Gaussian Process Implicit Surfaces presented at RA-L and ICRA2021.

The source code is modified from Bhoaram Lee's GPisMap code.

License

Licensed under GNU General Public License version 3.

Requirements: Software

  1. Eigen

  2. MATLAB

Compiling and Running

  1. Clone this repository
git clone https://github.com/lanwu076/LogGPisMap.git
  1. Go into the mex directory in MATLAB
cd mex
  1. Compile the mex functions by executing the make script.

    • Setup mex
    mex -setup
    
    • Run the make scripts
    make_GPisMap
    
  2. Run the demo scripts

    • For 2D
    run('../matlab/demo_gpisMap.m')
    
  3. Trouble shooting

    • If mex complains about not finding eigen, configure the eigen path appropriately in make_GPisMap.m

Video

A demonstration video is available here.

Contributors

The major contributors of this work are Lan Wu, Ki Myung Brian Lee.

Citation

If you find LogGPisMap useful in your research, please consider citing:

@ARTICLE{lwu21-ral,
  author={Wu, Lan and Lee, Ki Myung Brian and Liu, Liyang and Vidal-Calleja, Teresa},
  journal={Rob. and Automat. Lett.}, 
  title={Faithful Euclidean Distance Field From Log-Gaussian Process Implicit Surfaces}, 
  year={2021},
  volume={6},
  number={2},
  pages={2461-2468}
}

About

Faithful Euclidean distance field from Log-Gaussian Process Implicit Surfaces

License:GNU General Public License v3.0


Languages

Language:C++ 89.3%Language:MATLAB 8.7%Language:C 2.0%