Detailed Description

Sparse solver for GMRF (Gaussian Markov Random Fields) graphical models.

The design of this class is optimized for large problems (e.g. >1e3 nodes, >1e4 constrainst) by leaving to the user/caller the responsibility of allocating all "nodes" and constraints. This class can be seen as an intermediary solution between current methods in mrpt::graphslam and the well-known G2O library:

Assumptions/limitations:

Linear error functions (for now).
Scalar (1-dim) error functions.
Gaussian factors.
Solver: Eigen SparseQR.

Usage:

Call initialize() to set the number of nodes.
Call addConstraints() to insert constraints. This may be called more than once.
Call updateEstimation() to run one step of the linear SparseQR solver.

Note: [New in MRPT 1.5.0] Requires Eigen>=3.1

Definition at line 44 of file ScalarFactorGraph.h.

#include <mrpt/graphs/ScalarFactorGraph.h>

Inheritance diagram for mrpt::graphs::ScalarFactorGraph:

Classes
struct	BinaryFactorVirtualBase
	Simple, scalar (1-dim) constraint (edge) for a GMRF. More...

struct	FactorBase

struct	UnaryFactorVirtualBase
	Simple, scalar (1-dim) constraint (edge) for a GMRF. More...

Public Member Functions
	ScalarFactorGraph ()

void	clear ()
	Reset state: remove all constraints and nodes. More...

void	initialize (const size_t nodeCount)
	Initialize the GMRF internal state and copy the prior factors. More...

void	addConstraint (const UnaryFactorVirtualBase &listOfConstraints)
	Insert constraints into the GMRF problem. More...

void	addConstraint (const BinaryFactorVirtualBase &listOfConstraints)

bool	eraseConstraint (const FactorBase &c)
	Removes a constraint. More...

void	clearAllConstraintsByType_Unary ()

void	clearAllConstraintsByType_Binary ()

void	updateEstimation (Eigen::VectorXd &solved_x_inc, Eigen::VectorXd *solved_variances=nullptr)

bool	isProfilerEnabled () const

void	enableProfiler (bool enable=true)

Private Attributes
size_t	m_numNodes
	number of nodes in the graph More...

std::deque< const UnaryFactorVirtualBase * >	m_factors_unary

std::deque< const BinaryFactorVirtualBase * >	m_factors_binary

mrpt::utils::CTimeLogger	m_timelogger

bool	m_enable_profiler

Constructor & Destructor Documentation

◆ ScalarFactorGraph()

ScalarFactorGraph::ScalarFactorGraph ( )

Definition at line 26 of file ScalarFactorGraph.cpp.

Member Function Documentation

◆ addConstraint() [1/2]

void ScalarFactorGraph::addConstraint ( const UnaryFactorVirtualBase & listOfConstraints )

Insert constraints into the GMRF problem.

Parameters

listOfConstraints List of user-implemented constraints. A pointer to the passed object is kept, but memory ownship REMAINS being responsability of the caller. This is done such that arrays/vectors of constraints can be more efficiently allocated if their type is known at build time.

Definition at line 47 of file ScalarFactorGraph.cpp.

References m_factors_unary.

Referenced by mrpt::maps::CRandomFieldGridMap3D::insertIndividualReading(), and mrpt::maps::CRandomFieldGridMap2D::insertObservation_GMRF().

Here is the caller graph for this function:

◆ addConstraint() [2/2]

void ScalarFactorGraph::addConstraint ( const BinaryFactorVirtualBase & listOfConstraints )

Definition at line 51 of file ScalarFactorGraph.cpp.

References m_factors_binary.

◆ clear()

void ScalarFactorGraph::clear ( )

Reset state: remove all constraints and nodes.

Definition at line 31 of file ScalarFactorGraph.cpp.

References m_factors_binary, m_factors_unary, m_numNodes, and MRPT_LOG_DEBUG.

◆ clearAllConstraintsByType_Binary()

void mrpt::graphs::ScalarFactorGraph::clearAllConstraintsByType_Binary ( )

inline

Definition at line 96 of file ScalarFactorGraph.h.

References m_factors_binary.

◆ clearAllConstraintsByType_Unary()

void mrpt::graphs::ScalarFactorGraph::clearAllConstraintsByType_Unary ( )

inline

Definition at line 95 of file ScalarFactorGraph.h.

References m_factors_unary.

◆ enableProfiler()

void mrpt::graphs::ScalarFactorGraph::enableProfiler ( bool enable = true )

inline

Definition at line 105 of file ScalarFactorGraph.h.

References m_enable_profiler.

Referenced by mrpt::maps::CRandomFieldGridMap2D::enableProfiler().

Here is the caller graph for this function:

◆ eraseConstraint()

bool ScalarFactorGraph::eraseConstraint ( const FactorBase & c )

Removes a constraint.

Return true if found and deleted correctly.

Definition at line 56 of file ScalarFactorGraph.cpp.

References mrpt::utils::find(), m_factors_binary, and m_factors_unary.

Referenced by mrpt::maps::CRandomFieldGridMap2D::updateMapEstimation_GMRF().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ initialize()

void ScalarFactorGraph::initialize ( const size_t nodeCount )

Initialize the GMRF internal state and copy the prior factors.

Parameters

nodeCount Number of unknown nodes in the MRF graph

Definition at line 40 of file ScalarFactorGraph.cpp.

References m_numNodes, and MRPT_LOG_DEBUG_STREAM.

◆ isProfilerEnabled()

bool mrpt::graphs::ScalarFactorGraph::isProfilerEnabled ( ) const

inline

Definition at line 104 of file ScalarFactorGraph.h.

References m_enable_profiler.

Referenced by mrpt::maps::CRandomFieldGridMap2D::isProfilerEnabled().

Here is the caller graph for this function:

◆ updateEstimation()

void ScalarFactorGraph::updateEstimation	(	Eigen::VectorXd &	solved_x_inc,
		Eigen::VectorXd *	solved_variances = `nullptr`
	)

Parameters

solved_x_inc	Output increment of the current estimate. Caller must add this vector to current state vector to obtain the optimal estimation.
solved_variances	If !=nullptr, the variances of each estimate will be stored here.

Definition at line 85 of file ScalarFactorGraph.cpp.

References ASSERT_, ASSERTMSG_, mrpt::utils::CTimeLogger::enable(), mrpt::utils::CTimeLogger::enter(), mrpt::utils::CTimeLogger::leave(), m_enable_profiler, m_factors_binary, m_factors_unary, m_numNodes, m_timelogger, MRPT_LOG_DEBUG_FMT, MRPT_TODO, mrpt::math::sum(), and THROW_EXCEPTION.