Class CObservation2DRangeScanWithUncertainty

Nested Relationships

Nested Types

Class Documentation

class CObservation2DRangeScanWithUncertainty

A 2D range scan plus an uncertainty model for each range.

Public Functions

double evaluateScanLikelihood(const CObservation2DRangeScan &otherScan, const TEvalParams &params) const

Returns a measure of the likelihood of a given scan, compared to this scan variances

Public Members

CObservation2DRangeScan rangeScan

The observation with the mean ranges in the scan field

mrpt::math::CVectorDouble rangesMean

The same ranges than in rangeScan.getScanRange(), for convenience as a math vector container, and with double precision

mrpt::math::CMatrixDouble rangesCovar

The covariance matrix for all the ranges in rangeScan.getScanRange()

struct TEvalParams

Public Functions

TEvalParams()

Public Members

double prob_outliers = {0.5}

(Default: 0.5) Probability of having an outlier (dynamic obstacles, not mapped) in each scan ray.

double prob_lost_ray = {0.3}

(Default: 0.3) Conditional probability: how many of the “no return” ranges come from a failure to detect a real obstacle.

double max_prediction_std_dev = {1.0}

(Default: 1.0m) Maximum std deviation of overall uncertainty for a range prediction to be considered as reliable for evaluation

double min_ray_log_lik = {-20.0}

(Default: -20) Minimum log-likelihood of a single ray