Struct RelocalizationLikelihood_SE2

• Defined in File relocalization.h

Nested Relationships

Nested Types

• Struct RelocalizationLikelihood_SE2::Input

• Struct RelocalizationLikelihood_SE2::Output

Struct Documentation

struct RelocalizationLikelihood_SE2

Takes a global metric map, an observation, and a SE(2) ROI, and evaluates the likelihood of the observation in a regular SE(2) lattice.

This is based on mrpt maps observationLikelihood() evaluation, so the main parameters that determine the way likelihood is computed must be defined on before hand within each of the metric map layers of the input reference map.

At present this algorithm is useful for these sensor/map types:

• observations: pointclouds/2d_scan, reference_map: 2D gridmap

• observations: pointclouds, reference_map: pointclouds

  Todo:

  Implement likelihood of observations GNNS datum within georeferenced maps.

Public Static Functions

static Output run(const Input &in)

struct Input

Public Functions

Input() = default

Public Members

mp2p_icp::metric_map_t reference_map

mrpt::obs::CSensoryFrame observations

mrpt::math::TPose2D corner_min

mrpt::math::TPose2D corner_max

double resolution_xy = 0.5

double resolution_phi = mrpt::DEG2RAD(30.0)

struct Output

Public Functions

Output() = default

Public Members

mrpt::poses::CPosePDFGrid likelihood_grid

double time_cost = .0

[s]

double max_log_likelihood = 0

double min_log_likelihood = 0