#Introduction

---

This repository contains the accompanying code of our IROS'12 paper Fast Minimum Uncertainty Search on a Graph Map Representation.

In the paper we takle the problem of how to plan the minimum uncertainty path in a roadmap like structure?.

We have proposed a fast path planning algorithm capable of obtaining the minimum uncertainty path according to a reduced representation of the environment using a determinant-based criterion (e.g. D-opt).

The code in this repository is based on the code used in the experiments report in the paper. Moreover, It can be used as a starting point in an minimum uncertainty navigation framework based on SLAM algorithms.

If you use this work, please cite our corresponding paper :

@INPROCEEDINGS{Carrillo2012,
author = {H. Carrillo and Y. Latif and J. Neira and J. A. Castellanos},
title = {{Fast Minimum Uncertainty Search on a Graph Map Representation}},
booktitle = {IEEE / RSJ International Conference on Intelligent Robots and Systems, IROS’12},
year = {2012},
address = {Vilamoura, Algarve, Portugal},
month = {October}
}

#Requirements

---

• Suitesparse ( Avalible in http://www.cise.ufl.edu/research/sparse/SuiteSparse/ Version 3.4.0 )
  • Also with sudo apt-get install libsuitesparse-dev
• g2o (avalible in http://openslam.org/g2o Rev. 30)
  • If you get "No rule to make target `/usr/lib/x86_64-linux-gnu/libGL.so'" (probably becasue you are using 64 bits linux and a NVIDIA card). Check the symbolic link of libGL.so and point it to the correct place.
• Eigen3 (avalible in http://eigen.tuxfamily.org/index.php?title=Main_Page Version 3.1.1)
• Nanoflann (avalible in http://code.google.com/p/nanoflann/ Version 1.1.3)
• Boost graph 1.42 (avalible in http://www.boost.org/)
  • Version 1.42 is strongly recommended

Although g2o was used for the experiment, as stated in the paper others graph-SLAM based implementations could be used (e.g. iSAM), as long as the correct interface is coded.

#Installation

---

• Install all the requirements.
• Go to the build folder
  • cd build
• compile tesp.cpp
  • make

The code was tested under Kubuntu 10.04 32 bits.

#Usage

---

• Usage: ./FaMuS g2o_filename startVertexID endVertexID
  -./FaMuS ../Data/intel_opt.g2o 10 500
  The outputs in the command line are:

• The nodes of the shortest path and the minimum uncertainty path.
• The percentage of overlap between the shortest path and the minimum uncertainty path.
• Overall time of the reduction process.
• Final uncertainty value of the shortest path.

Also if WRITE_MATLAB is defined, a MATLAB script with the resulting paths will be generated. In the folder named "Results" is available a MATLAB script (i.e. plottingFunctions.m) to plot figures presented in the paper.

#Contact

---

If you have any queries please contact:

Henry Carrillo L.
http://webdiis.unizar.es/~hcarri/

Yasir Latif (ylatif at unizar dot es)