Welcome to TORO - Tree-based netwORk Optimizer

What is TORO?

TORO is an optimization approach for constraint-network. It provides a highly efficient, gradient descent-based error minimization procedure. Currently (2007), there are not many such efficient methods available (at least we have not seen them yet...) There is a 2D and a 3D version of TORO available.

In 2006, Olson et al. presented a novel approach to solve the graph-based SLAM problem by applying stochastic gradient descent to minimize the error introduced by constraints. TORO is an extension of Olson's algorithm. It applies a tree parameterization of the nodes in the graph that significantly improves the performance and enables a robot to cope with arbitrary network topologies. The latter allows us to bound the complexity of the algorithm to the size of the mapped area and not to the length of the trajectory.

Update 11/2007: We provide a new version of TORO that works also in 3D and not only in 2D.

Copyright and License Information

Copyright, 2007 by Giorgio Grisetti, Cyrill Stachniss, and Wolfram Burgard

This software is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

TORO is licences under the Common Creative License, Attribution-NonCommercial-ShareAlike 3.0

You are free:

• to Share - to copy, distribute and transmit the work
• to Remix - to adapt the work

Under the following conditions:

• Attribution. You must attribute the work in the manner specified by the author or licensor (but not in any way that suggests that they endorse you or your use of the work).
• Noncommercial. You may not use this work for commercial purposes.
• Share Alike. If you alter, transform, or build upon this work, you may distribute the resulting work only under the same or similar license to this one.

Any of the above conditions can be waived if you get permission from the copyright holder. Nothing in this license impairs or restricts the author's moral rights.

Quick Installation Guide (from an archive)

$ tar xzf toro.tgz ~/
$ cd toro
$ make

Quick Installation Guide (from the repositroy)

$ svn co http://www.openslam.org/data/svn/toro
$ cd toro
$ make

Quick Guide: Running TORO on an example file

$ ./toro examples/simple.graph
$ gnuplot XXXXX

Long Guide: Running TORO (2D)

You have the following command line options

$ ./toro [options] <graph-file>
 Options are 
 -vl <int>  set verbosity level
 -dr        disables node reduction
 -mst       use the minimum spanning tree
 -ic        enable index renaming (saves memory)
 -nib       disable initialization according to observations
 -i <int>   perform <int> iterations
 -df        dump gnuplot files of the intermediateresults
 -nde       disable dump error on screen (saves time)
 -oc        override covariances from the file (deals with corrupted input)

Long Guide: Running TORO3D

You have the following command line options

$ ./toro3d [options] <graph-file>
 Options are 
 -vl <int>  set verbosity level
 -dr        disables node reduction
 -mst       use the minimum spanning tree
 -ic        enable index renaming (saves memory)
 -nib       disable initialization according to observations
 -i <int>   perform <int> iterations
 -df        dump gnuplot files of the intermediate results
 -nde       disable dump error on screen (saves time)
 -ar        adaptive restart
 -oc        overrides covariances
 -ip        ignores preconditioner
 -2d        load a 2d file
 -sl        sort the constraints according to the length of the path
            instead that according to the level of a constraint

Further Reading

TORO3D is an implementation of the approach described in:

Grisetti Giorgio, Slawomir Grzonka, Cyrill Stachniss, Patrick Pfaff, and Wolfram Burgard Efficient Estimation of Accurate Maximum Likelihood Maps in 3D. In Proc. of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2007.

TORO (2D) is an implementation of the approach described in:

Giorgio Grisetti, Cyrill Stachniss, Slawomir Grzonka, and Wolfram Burgard A Tree Parameterization for Efficiently Computing Maximum Likelihood Maps using Gradient Descent. Robotics: Science and Systems (RSS), Atlanta, GA, USA, 2007.

It is highly recommended to read:

E. Olson, J.J. Leonard, and S. Teller. Fast iterative optimization of pose graphs with poor initial estimates. In Proc. of the IEEE Int. Conf. on Robotics & Automation (ICRA), pages 2262-2269, 2006.

---