Function mrpt::tfest::se2_l2_robust

Defined in File se2.h

Function Documentation

bool mrpt::tfest::se2_l2_robust(const mrpt::tfest::TMatchingPairList &in_correspondences, const double in_normalizationStd, const TSE2RobustParams &in_ransac_params, TSE2RobustResult &out_results)

Robust least-squares (L2 norm) solution to finding the optimal SE(2) (x,y,yaw) between two reference frames. This method implements a RANSAC-based robust estimation, returning a probability distribution over all the possibilities as a Sum of Gaussians.

The optimal transformation q fulfills \( p_{this} = q \oplus p_{other} \), that is, the transformation of frame other with respect to this.

../../output_staging/generated/doxygen/xml/tfest_frames.png

The technique was described in the paper:

J.L. Blanco, J. Gonzalez-Jimenez, J.A. Fernandez-Madrigal, “A Robust,

Multi-Hypothesis Approach to Matching Occupancy Grid Maps”, Robotica, 2013.
http://dx.doi.org/10.1017/S0263574712000732

This works as follows:

Repeat “ransac_nSimulations” times:
- Randomly pick TWO correspondences from the set “in_correspondences”.
- Compute the associated rigid transformation.
- For “ransac_maxSetSize” randomly selected correspondences, test for “consensus” with the current group:
  - If if is compatible (ransac_mahalanobisDistanceThreshold), grow the “consensus set”
  - If not, do not add it.
For more details refer to the tutorial on scan matching methods. NOTE: Parameter ransac_maxSetSize should be set to in_correspondences.size() to make sure that every correspondence is tested for each random permutation.