GTSAM  4.0.2
C++ library for smoothing and mapping (SAM)
Classes | Public Member Functions | Public Attributes | List of all members
gtsam::ISAM2Result Struct Reference

#include <ISAM2Result.h>

Collaboration diagram for gtsam::ISAM2Result:
Collaboration graph
[legend]

Classes

struct  DetailedResults
 

Public Member Functions

 ISAM2Result (bool enableDetailedResults=false)
 
DetailedResultsdetails ()
 Return pointer to detail, 0 if no detail requested.
 
void print (const std::string str="") const
 Print results.
 
size_t getVariablesRelinearized () const
 
size_t getVariablesReeliminated () const
 
FactorIndices getNewFactorsIndices () const
 
size_t getCliques () const
 
double getErrorBefore () const
 
double getErrorAfter () const
 

Public Attributes

std::optional< double > errorBefore
 
std::optional< double > errorAfter
 
size_t variablesRelinearized
 
size_t variablesReeliminated
 
size_t factorsRecalculated
 
size_t cliques
 
FactorIndices newFactorsIndices
 
KeySet unusedKeys
 
KeyVector observedKeys
 
KeySet keysWithRemovedFactors
 
KeySet markedKeys
 
std::optional< DetailedResultsdetail
 

Detailed Description

This struct is returned from ISAM2::update() and contains information about the update that is useful for determining whether the solution is converging, and about how much work was required for the update. See member variables for details and information about each entry.

Member Function Documentation

◆ getVariablesRelinearized()

size_t gtsam::ISAM2Result::getVariablesRelinearized ( ) const
inline

Getters and Setters

Member Data Documentation

◆ cliques

size_t gtsam::ISAM2Result::cliques

The number of cliques in the Bayes' Tree

◆ detail

std::optional<DetailedResults> gtsam::ISAM2Result::detail

Detailed results, if enabled by ISAM2Params::enableDetailedResults. See Detail for information about the results data stored here.

◆ errorAfter

std::optional<double> gtsam::ISAM2Result::errorAfter

The nonlinear error of all of the factors computed after the current update, meaning that variables above the relinearization threshold (ISAM2Params::relinearizeThreshold) have been relinearized and new variables have undergone one linear update. Variable values are again computed by combining their linearization points with their partial linear deltas, by ISAM2::calculateEstimate().

Note: This will only be computed if
ISAM2Params::evaluateNonlinearError is set to true, because there is some cost to this computation.

◆ errorBefore

std::optional<double> gtsam::ISAM2Result::errorBefore

The nonlinear error of all of the factors, including new factors and variables added during the current call to ISAM2::update(). This error is calculated using the following variable values:

  • Pre-existing variables will be evaluated by combining their linearization point before this call to update, with their partial linear delta, as computed by ISAM2::calculateEstimate().
  • New variables will be evaluated at their initialization points passed into the current call to update.
    Note: This will only be computed if
    ISAM2Params::evaluateNonlinearError is set to true, because there is some cost to this computation.

◆ factorsRecalculated

size_t gtsam::ISAM2Result::factorsRecalculated

The number of factors that were included in reelimination of the Bayes' tree.

◆ keysWithRemovedFactors

KeySet gtsam::ISAM2Result::keysWithRemovedFactors

Keys of variables that had factors removed.

◆ markedKeys

KeySet gtsam::ISAM2Result::markedKeys

All keys that were marked during the update process.

◆ newFactorsIndices

FactorIndices gtsam::ISAM2Result::newFactorsIndices

The indices of the newly-added factors, in 1-to-1 correspondence with the factors passed as newFactors to ISAM2::update(). These indices may be used later to refer to the factors in order to remove them.

◆ observedKeys

KeyVector gtsam::ISAM2Result::observedKeys

keys for variables that were observed, i.e., not unused.

◆ unusedKeys

KeySet gtsam::ISAM2Result::unusedKeys

Unused keys, and indices for unused keys, i.e., keys that are empty now and do not appear in the new factors.

◆ variablesReeliminated

size_t gtsam::ISAM2Result::variablesReeliminated

The number of variables that were reeliminated as parts of the Bayes' Tree were recalculated, due to new factors. When loop closures occur, this count will be large as the new loop-closing factors will tend to involve variables far away from the root, and everything up to the root will be reeliminated.

◆ variablesRelinearized

size_t gtsam::ISAM2Result::variablesRelinearized

The number of variables that were relinearized because their linear deltas exceeded the reslinearization threshold (ISAM2Params::relinearizeThreshold), combined with any additional variables that had to be relinearized because they were involved in the same factor as a variable above the relinearization threshold. On steps where no relinearization is considered (see ISAM2Params::relinearizeSkip), this count will be zero.


The documentation for this struct was generated from the following file: