Detailed Description

Declares a class that represents a Rao-Blackwellized set of particles for solving the SLAM problem (This class is the base of RBPF-SLAM applications).

This class is used internally by the map building algorithm in "mrpt::slam::CMetricMapBuilderRBPF"

See also:: mrpt::slam::CMetricMapBuilderRBPF

#include <mrpt/slam/CMultiMetricMapPDF.h>

Inheritance diagram for mrpt::slam::CMultiMetricMapPDF:

[legend]

List of all members.

Classes
struct	TPredictionParams
	The struct for passing extra simulation parameters to the prediction/update stage when running a particle filter. More...
Public Types
typedef T	CParticleDataContent
	This is the type inside the corresponding CParticleData class.
typedef CProbabilityParticle< T >	CParticleData
	Use this to refer to each element in the m_particles array.
typedef std::deque< CParticleData >	CParticleList
	Use this type to refer to the list of particles m_particles.
typedef double(*	TParticleProbabilityEvaluator )(const bayes::CParticleFilter::TParticleFilterOptions &PF_options, const CParticleFilterCapable obj, size_t index, const void action, const void *observation)
	A callback function type for evaluating the probability of m_particles of being selected, used in "fastDrawSample".
Public Member Functions
	IMPLEMENT_PARTICLE_FILTER_CAPABLE (CRBPFParticleData)
	CMultiMetricMapPDF (const bayes::CParticleFilter::TParticleFilterOptions &opts=bayes::CParticleFilter::TParticleFilterOptions(), const mrpt::slam::TSetOfMetricMapInitializers mapsInitializers=NULL, const TPredictionParams predictionOptions=NULL)
	Constructor.
virtual	~CMultiMetricMapPDF ()
	Destructor.
void	clear (const CPose2D &initialPose)
	Clear all elements of the maps, and restore all paths to a single starting pose.
void	clear (const CPose3D &initialPose)
	Clear all elements of the maps, and restore all paths to a single starting pose.
void	getEstimatedPosePDFAtTime (size_t timeStep, CPose3DPDFParticles &out_estimation) const
	Returns the estimate of the robot pose as a particles PDF for the instant of time "timeStep", from 0 to N-1.
void	getEstimatedPosePDF (CPose3DPDFParticles &out_estimation) const
	Returns the current estimate of the robot pose, as a particles PDF.
CMultiMetricMap *	getCurrentMetricMapEstimation ()
	Returns the weighted averaged map based on the current best estimation.
CMultiMetricMap *	getCurrentMostLikelyMetricMap ()
	Returns a pointer to the current most likely map (associated to the most likely particle).
size_t	getNumberOfObservationsInSimplemap () const
	Get the number of CSensoryFrame inserted into the internal member SFs.
void	insertObservation (CSensoryFrame &sf)
	Insert an observation to the map, at each particle's pose and to each particle's metric map.
void	getPath (size_t i, std::deque< math::TPose3D > &out_path) const
	Return the path (in absolute coordinate poses) for the i'th particle.
double	getCurrentEntropyOfPaths ()
	Returns the current entropy of paths, computed as the average entropy of poses along the path, where entropy of each pose estimation is computed as the entropy of the gaussian approximation covariance.
double	getCurrentJointEntropy ()
	Returns the joint entropy estimation over paths and maps, acording to "Information Gain-based Exploration Using" by C.
void	updateSensoryFrameSequence ()
	Update the poses estimation of the member "SFs" according to the current path belief.
void	saveCurrentPathEstimationToTextFile (const std::string &fil)
	A logging utility: saves the current path estimation for each particle in a text file (a row per particle, each 3-column-entry is a set [x,y,phi], respectively).
mrpt::utils::CObjectPtr	duplicateGetSmartPtr () const
	Returns a copy of the object, indepently of its class, as a smart pointer (the newly created object will exist as long as any copy of this smart pointer).
CObject *	clone () const
	Cloning interface for smart pointers.
void	clearParticles ()
	Free the memory of all the particles and reset the array "m_particles" to length zero.
void	writeParticlesToStream (utils::CStream &out) const
	Dumps the sequence of particles and their weights to a stream (requires T implementing CSerializable).
void	readParticlesFromStream (utils::CStream &in)
	Reads the sequence of particles and their weights from a stream (requires T implementing CSerializable).
void	getWeights (vector_double &out_logWeights) const
	Returns a vector with the sequence of the logaritmic weights of all the samples.
const CParticleData *	getMostLikelyParticle () const
	Returns the particle with the highest weight.
void	prepareFastDrawSample (const bayes::CParticleFilter::TParticleFilterOptions &PF_options, TParticleProbabilityEvaluator partEvaluator=defaultEvaluator, const void action=NULL, const void observation=NULL) const
	Prepares data structures for calling fastDrawSample method next.
size_t	fastDrawSample (const bayes::CParticleFilter::TParticleFilterOptions &PF_options) const
	Draws a random sample from the particle filter, in such a way that each particle has a probability proportional to its weight (in the standard PF algorithm).
virtual double	getW (size_t i) const =0
	Access to i'th particle (logarithm) weight, where first one is index 0.
virtual void	setW (size_t i, double w)=0
	Modifies i'th particle (logarithm) weight, where first one is index 0.
virtual size_t	particlesCount () const =0
	Get the m_particles count.
void	prediction_and_update (const mrpt::slam::CActionCollection action, const mrpt::slam::CSensoryFrame observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
	Performs the prediction stage of the Particle Filter.
virtual void	performSubstitution (const std::vector< size_t > &indx)=0
	Performs the substitution for internal use of resample in particle filter algorithm, don't call it directly.
virtual double	normalizeWeights (double *out_max_log_w=NULL)=0
	Normalize the (logarithmic) weights, such as the maximum weight is zero.
virtual double	ESS ()=0
	Returns the normalized ESS (Estimated Sample Size), in the range [0,1].
void	performResampling (const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
	Performs a resample of the m_particles, using the method selected in the constructor.
bool	PF_SLAM_implementation_gatherActionsCheckBothActObs (const CActionCollection actions, const CSensoryFrame sf)
	Auxiliary method called by PF implementations: return true if we have both action & observation, otherwise, return false AND accumulate the odometry so when we have an observation we didn't lose a thing.
Virtual methods that the PF_implementations assume exist.
const TPose3D *	getLastPose (const size_t i) const
	Return a pointer to the last robot pose in the i'th particle (or NULL if it's a path and it's empty).
void	PF_SLAM_implementation_custom_update_particle_with_new_pose (CParticleDataContent *particleData, const TPose3D &newPose) const
bool	PF_SLAM_implementation_doWeHaveValidObservations (const CParticleList &particles, const CSensoryFrame *sf) const
bool	PF_SLAM_implementation_skipRobotMovement () const
	Make a specialization if needed, eg.
double	PF_SLAM_computeObservationLikelihoodForParticle (const CParticleFilter::TParticleFilterOptions &PF_options, const size_t particleIndexForMap, const CSensoryFrame &observation, const CPose3D &x) const
	Evaluate the observation likelihood for one particle at a given location.
Virtual methods that the PF_implementations assume exist.
virtual void	PF_SLAM_implementation_custom_update_particle_with_new_pose (PARTICLE_TYPE *particleData, const TPose3D &newPose) const =0
virtual void	PF_SLAM_implementation_replaceByNewParticleSet (typename CParticleFilterData< PARTICLE_TYPE >::CParticleList &old_particles, const vector< TPose3D > &newParticles, const vector_double &newParticlesWeight, const vector< size_t > &newParticlesDerivedFromIdx) const
	This is the default algorithm to efficiently replace one old set of samples by another new set.
virtual bool	PF_SLAM_implementation_doWeHaveValidObservations (const typename CParticleFilterData< PARTICLE_TYPE >::CParticleList &particles, const CSensoryFrame *sf) const
Static Public Member Functions
static double	defaultEvaluator (const bayes::CParticleFilter::TParticleFilterOptions &PF_options, const CParticleFilterCapable obj, size_t index, const void action, const void *observation)
	The default evaluator function, which simply returns the particle weight.
static void	computeResampling (CParticleFilter::TParticleResamplingAlgorithm method, const vector_double &in_logWeights, std::vector< size_t > &out_indexes)
	A static method to perform the computation of the samples resulting from resampling a given set of particles, given their logarithmic weights, and a resampling method.
static void	log2linearWeights (const vector_double &in_logWeights, vector_double &out_linWeights)
	A static method to compute the linear, normalized (the sum the unity) weights from log-weights.
Public Attributes
mrpt::slam::CMultiMetricMapPDF::TPredictionParams	options
float	newInfoIndex
	An index [0,1] measuring how much information an observation aports to the map (Typ.
CParticleList	m_particles
	The array of particles.
Static Public Attributes
static const mrpt::utils::TRuntimeClassId	classCObject
Protected Member Functions
void	prediction_and_update_pfStandardProposal (const mrpt::slam::CActionCollection action, const mrpt::slam::CSensoryFrame observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
	The PF algorithm implementation.
void	prediction_and_update_pfOptimalProposal (const mrpt::slam::CActionCollection action, const mrpt::slam::CSensoryFrame observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
	The PF algorithm implementation.
void	prediction_and_update_pfAuxiliaryPFOptimal (const mrpt::slam::CActionCollection action, const mrpt::slam::CSensoryFrame observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
	The PF algorithm implementation.
virtual void	writeToStream (mrpt::utils::CStream &out, int *getVersion) const =0
	Introduces a pure virtual method responsible for writing to a CStream.
virtual void	readFromStream (mrpt::utils::CStream &in, int version)=0
	Introduces a pure virtual method responsible for loading from a CStream This can not be used directly be users, instead use "stream >> object;" for reading it from a stream or "stream >> object_ptr;" if the class is unknown apriori.
virtual void	prediction_and_update_pfAuxiliaryPFStandard (const mrpt::slam::CActionCollection action, const mrpt::slam::CSensoryFrame observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
	Performs the particle filter prediction/update stages for the algorithm "pfAuxiliaryPFStandard" (if not implemented in heritated class, it will raise a 'non-implemented' exception).
The generic PF implementations for localization & SLAM.
void	PF_SLAM_implementation_pfAuxiliaryPFOptimal (const CActionCollection actions, const CSensoryFrame sf, const CParticleFilter::TParticleFilterOptions &PF_options, const TKLDParams &KLD_options)
	A generic implementation of the PF method "prediction_and_update_pfAuxiliaryPFOptimal" (optimal sampling with rejection sampling approximation), common to both localization and mapping.
void	PF_SLAM_implementation_pfAuxiliaryPFStandard (const CActionCollection actions, const CSensoryFrame sf, const CParticleFilter::TParticleFilterOptions &PF_options, const TKLDParams &KLD_options)
	A generic implementation of the PF method "prediction_and_update_pfAuxiliaryPFStandard" (Auxiliary particle filter with the standard proposal), common to both localization and mapping.
void	PF_SLAM_implementation_pfStandardProposal (const CActionCollection actions, const CSensoryFrame sf, const CParticleFilter::TParticleFilterOptions &PF_options, const TKLDParams &KLD_options)
	A generic implementation of the PF method "pfStandardProposal" (standard proposal distribution, that is, a simple SIS particle filter), common to both localization and mapping.
Protected Attributes
TFastDrawAuxVars	m_fastDrawAuxiliary
	Auxiliary vectors, see CParticleFilterCapable::prepareFastDrawSample for more information.
Private Member Functions
float	H (float p)
	Entropy aux.
void	rebuildAverageMap ()
	Rebuild the "expected" grid map.
Private Attributes
CMultiMetricMap	averageMap
	Internal buffer for the averaged map.
bool	averageMapIsUpdated
CSimpleMap	SFs
	The SFs and their corresponding pose estimations:
std::vector< uint32_t >	SF2robotPath
	A mapping between indexes in the SFs to indexes in the robot paths from particles.
Friends
class	CMetricMapBuilderRBPF
RTTI stuff
class	mrpt::utils::CStream
static const mrpt::utils::TRuntimeClassId	classCSerializable
RTTI stuff
typedef CMultiMetricMapPDFPtr	SmartPtr
static mrpt::utils::CLASSINIT	_init_CMultiMetricMapPDF
static mrpt::utils::TRuntimeClassId	classCMultiMetricMapPDF
static const mrpt::utils::TRuntimeClassId *	classinfo
static const mrpt::utils::TRuntimeClassId *	_GetBaseClass ()
virtual const mrpt::utils::TRuntimeClassId *	GetRuntimeClass () const
	Returns information about the class of an object in runtime.
virtual mrpt::utils::CObject *	duplicate () const
	Returns a copy of the object, indepently of its class.
static mrpt::utils::CObject *	CreateObject ()
static CMultiMetricMapPDFPtr	Create ()
Data members and methods used by generic PF implementations
static double	PF_SLAM_particlesEvaluator_AuxPFStandard (const CParticleFilter::TParticleFilterOptions &PF_options, const CParticleFilterCapable obj, size_t index, const void action, const void *observation)
	Compute w[i]·p(z_t \| mu_t^i), with mu_t^i being the mean of the new robot pose.
static double	PF_SLAM_particlesEvaluator_AuxPFOptimal (const CParticleFilter::TParticleFilterOptions &PF_options, const CParticleFilterCapable obj, size_t index, const void action, const void *observation)
CActionRobotMovement2D	m_accumRobotMovement2D
bool	m_accumRobotMovement2DIsValid
CPose3DPDFGaussian	m_accumRobotMovement3D
bool	m_accumRobotMovement3DIsValid
CPoseRandomSampler	m_movementDrawer
	Used in al PF implementations.
vector_double	m_pfAuxiliaryPFOptimal_estimatedProb
	Auxiliary variable used in the "pfAuxiliaryPFOptimal" algorithm.
vector_double	m_pfAuxiliaryPFStandard_estimatedProb
	Auxiliary variable used in the "pfAuxiliaryPFStandard" algorithm.
vector_double	m_pfAuxiliaryPFOptimal_maxLikelihood
	Auxiliary variable used in the "pfAuxiliaryPFOptimal" algorithm.
std::vector< TPose3D >	m_pfAuxiliaryPFOptimal_maxLikDrawnMovement
	Auxiliary variable used in the "pfAuxiliaryPFOptimal" algorithm.
std::vector< bool >	m_pfAuxiliaryPFOptimal_maxLikMovementDrawHasBeenUsed

Member Typedef Documentation

typedef CProbabilityParticle<T> mrpt::bayes::CParticleFilterData::CParticleData [inherited]

Use this to refer to each element in the m_particles array.

Definition at line 61 of file CParticleFilterData.h.

typedef T mrpt::bayes::CParticleFilterData::CParticleDataContent [inherited]

This is the type inside the corresponding CParticleData class.

Definition at line 60 of file CParticleFilterData.h.

typedef std::deque<CParticleData> mrpt::bayes::CParticleFilterData::CParticleList [inherited]

Use this type to refer to the list of particles m_particles.

Definition at line 62 of file CParticleFilterData.h.

typedef CMultiMetricMapPDFPtr mrpt::slam::CMultiMetricMapPDF::SmartPtr

A typedef for the associated smart pointer

Definition at line 89 of file CMultiMetricMapPDF.h.

typedef double( * mrpt::bayes::CParticleFilterCapable::TParticleProbabilityEvaluator)(const bayes::CParticleFilter::TParticleFilterOptions &PF_options, const CParticleFilterCapable *obj, size_t index, const void *action, const void *observation) [inherited]

A callback function type for evaluating the probability of m_particles of being selected, used in "fastDrawSample".

The default evaluator function "defaultEvaluator" simply returns the particle weight.

Parameters:

index	This is the index of the particle its probability is being computed.
action	The value of this is the parameter passed to "prepareFastDrawSample"
observation	The value of this is the parameter passed to "prepareFastDrawSample" The action and the observation are declared as "void*" for a greater flexibility.

See also:: prepareFastDrawSample

Definition at line 73 of file CParticleFilterCapable.h.

Constructor & Destructor Documentation

mrpt::slam::CMultiMetricMapPDF::CMultiMetricMapPDF	(	const bayes::CParticleFilter::TParticleFilterOptions &	opts = `bayes::CParticleFilter::TParticleFilterOptions()`,
		const mrpt::slam::TSetOfMetricMapInitializers *	mapsInitializers = `NULL`,
		const TPredictionParams *	predictionOptions = `NULL`
	)

Constructor.

virtual mrpt::slam::CMultiMetricMapPDF::~CMultiMetricMapPDF ( ) [virtual]

Destructor.

Member Function Documentation

static const mrpt::utils::TRuntimeClassId* mrpt::slam::CMultiMetricMapPDF::_GetBaseClass ( ) [static, protected]

Reimplemented from mrpt::utils::CSerializable.

void mrpt::slam::CMultiMetricMapPDF::clear ( const CPose2D & initialPose )

Clear all elements of the maps, and restore all paths to a single starting pose.

void mrpt::slam::CMultiMetricMapPDF::clear ( const CPose3D & initialPose )

Clear all elements of the maps, and restore all paths to a single starting pose.

void mrpt::bayes::CParticleFilterData::clearParticles ( ) [inline, inherited]

Free the memory of all the particles and reset the array "m_particles" to length zero.

Definition at line 72 of file CParticleFilterData.h.

References MRPT_START, mrpt::bayes::CParticleFilterData::m_particles, and MRPT_END.

CObject* mrpt::utils::CObject::clone ( ) const [inline, inherited]

Cloning interface for smart pointers.

Reimplemented in mrpt::opengl::CRenderizable, and mrpt::opengl::CRenderizableDisplayList.

Definition at line 154 of file CObject.h.

static void mrpt::bayes::CParticleFilterCapable::computeResampling	(	CParticleFilter::TParticleResamplingAlgorithm	method,
		const vector_double &	in_logWeights,
		std::vector< size_t > &	out_indexes
	)		`[static, inherited]`

A static method to perform the computation of the samples resulting from resampling a given set of particles, given their logarithmic weights, and a resampling method.

It returns the sequence of indexes from the resampling. The number of output samples is the same than the input population. This generic method just computes these indexes, to actually perform a resampling in a particle filter object, call performResampling

See also:: performResampling

static CMultiMetricMapPDFPtr mrpt::slam::CMultiMetricMapPDF::Create ( ) [static]

static mrpt::utils::CObject* mrpt::slam::CMultiMetricMapPDF::CreateObject ( ) [static]

static double mrpt::bayes::CParticleFilterCapable::defaultEvaluator	(	const bayes::CParticleFilter::TParticleFilterOptions &	PF_options,
		const CParticleFilterCapable *	obj,
		size_t	index,
		const void *	action,
		const void *	observation
	)		`[inline, static, inherited]`

The default evaluator function, which simply returns the particle weight.

The action and the observation are declared as "void*" for a greater flexibility.

See also:: prepareFastDrawSample

Definition at line 84 of file CParticleFilterCapable.h.

References MRPT_UNUSED_PARAM, and mrpt::bayes::CParticleFilterCapable::getW().

virtual mrpt::utils::CObject* mrpt::slam::CMultiMetricMapPDF::duplicate ( ) const [virtual]

Returns a copy of the object, indepently of its class.

Implements mrpt::utils::CObject.

mrpt::utils::CObjectPtr mrpt::utils::CObject::duplicateGetSmartPtr ( ) const [inline, inherited]

Returns a copy of the object, indepently of its class, as a smart pointer (the newly created object will exist as long as any copy of this smart pointer).

Definition at line 151 of file CObject.h.

virtual double mrpt::bayes::CParticleFilterCapable::ESS ( ) [pure virtual, inherited]

Returns the normalized ESS (Estimated Sample Size), in the range [0,1].

Note that you do NOT need to normalize the weights before calling this.

size_t mrpt::bayes::CParticleFilterCapable::fastDrawSample ( const bayes::CParticleFilter::TParticleFilterOptions & PF_options ) const [inherited]

Draws a random sample from the particle filter, in such a way that each particle has a probability proportional to its weight (in the standard PF algorithm).

This method can be used to generate a variable number of m_particles when resampling: to vary the number of m_particles in the filter. See prepareFastDrawSample for more information, or the Particle Filter tutorial.

NOTES:

You MUST call "prepareFastDrawSample" ONCE before calling this method. That method must be called after modifying the particle filter (executing one step, resampling, etc...)
This method returns ONE index for the selected ("drawn") particle, in the range [0,M-1]
You do not need to call "normalizeWeights" before calling this.
See also:
prepareFastDrawSample

double mrpt::slam::CMultiMetricMapPDF::getCurrentEntropyOfPaths ( )

Returns the current entropy of paths, computed as the average entropy of poses along the path, where entropy of each pose estimation is computed as the entropy of the gaussian approximation covariance.

double mrpt::slam::CMultiMetricMapPDF::getCurrentJointEntropy ( )

Returns the joint entropy estimation over paths and maps, acording to "Information Gain-based Exploration Using" by C.

Stachniss, G. Grissetti and W.Burgard.

CMultiMetricMap* mrpt::slam::CMultiMetricMapPDF::getCurrentMetricMapEstimation ( )

Returns the weighted averaged map based on the current best estimation.

If you need a persistent copy of this object, please use "CSerializable::duplicate" and use the copy.

CMultiMetricMap* mrpt::slam::CMultiMetricMapPDF::getCurrentMostLikelyMetricMap ( )

Returns a pointer to the current most likely map (associated to the most likely particle).

void mrpt::slam::CMultiMetricMapPDF::getEstimatedPosePDF ( CPose3DPDFParticles & out_estimation ) const

Returns the current estimate of the robot pose, as a particles PDF.

See also:: getEstimatedPosePDFAtTime

void mrpt::slam::CMultiMetricMapPDF::getEstimatedPosePDFAtTime	(	size_t	timeStep,
		CPose3DPDFParticles &	out_estimation
	)		const

Returns the estimate of the robot pose as a particles PDF for the instant of time "timeStep", from 0 to N-1.

See also:: getEstimatedPosePDF

const TPose3D* mrpt::slam::CMultiMetricMapPDF::getLastPose ( const size_t i ) const [virtual]

Return a pointer to the last robot pose in the i'th particle (or NULL if it's a path and it's empty).

Implements mrpt::slam::PF_implementation< CRBPFParticleData, CMultiMetricMapPDF >.

const CParticleData* mrpt::bayes::CParticleFilterData::getMostLikelyParticle ( ) const [inline, inherited]

Returns the particle with the highest weight.

Definition at line 140 of file CParticleFilterData.h.

References MRPT_START, ASSERT_, mrpt::bayes::CParticleFilterData::m_particles, and MRPT_END.

size_t mrpt::slam::CMultiMetricMapPDF::getNumberOfObservationsInSimplemap ( ) const [inline]

Get the number of CSensoryFrame inserted into the internal member SFs.

Definition at line 222 of file CMultiMetricMapPDF.h.

void mrpt::slam::CMultiMetricMapPDF::getPath	(	size_t	i,
		std::deque< math::TPose3D > &	out_path
	)		const

Return the path (in absolute coordinate poses) for the i'th particle.

Exceptions:

On	index out of bounds

virtual const mrpt::utils::TRuntimeClassId* mrpt::slam::CMultiMetricMapPDF::GetRuntimeClass ( ) const [virtual]

Returns information about the class of an object in runtime.

Reimplemented from mrpt::utils::CSerializable.

virtual double mrpt::bayes::CParticleFilterCapable::getW ( size_t i ) const [pure virtual, inherited]

Access to i'th particle (logarithm) weight, where first one is index 0.

Referenced by mrpt::bayes::CParticleFilterCapable::defaultEvaluator().

void mrpt::bayes::CParticleFilterData::getWeights ( vector_double & out_logWeights ) const [inline, inherited]

Returns a vector with the sequence of the logaritmic weights of all the samples.

Definition at line 127 of file CParticleFilterData.h.

References MRPT_START, mrpt::dynamicsize_vector::resize(), mrpt::bayes::CParticleFilterData::m_particles, and MRPT_END.

float mrpt::slam::CMultiMetricMapPDF::H ( float p ) [private]

Entropy aux.

function

mrpt::slam::CMultiMetricMapPDF::IMPLEMENT_PARTICLE_FILTER_CAPABLE ( CRBPFParticleData )

void mrpt::slam::CMultiMetricMapPDF::insertObservation ( CSensoryFrame & sf )

Insert an observation to the map, at each particle's pose and to each particle's metric map.

Parameters:

sf	The SF to be inserted

static void mrpt::bayes::CParticleFilterCapable::log2linearWeights	(	const vector_double &	in_logWeights,
		vector_double &	out_linWeights
	)		`[static, inherited]`

A static method to compute the linear, normalized (the sum the unity) weights from log-weights.

See also:: performResampling

virtual double mrpt::bayes::CParticleFilterCapable::normalizeWeights ( double * out_max_log_w = NULL ) [pure virtual, inherited]

Normalize the (logarithmic) weights, such as the maximum weight is zero.

Parameters:

out_max_log_w If provided, will return with the maximum log_w before normalizing, such as new_weights = old_weights - max_log_w.

Returns:: The max/min ratio of weights ("dynamic range")

virtual size_t mrpt::bayes::CParticleFilterCapable::particlesCount ( ) const [pure virtual, inherited]

Get the m_particles count.

void mrpt::bayes::CParticleFilterCapable::performResampling ( const bayes::CParticleFilter::TParticleFilterOptions & PF_options ) [inherited]

Performs a resample of the m_particles, using the method selected in the constructor.

After computing the surviving samples, this method internally calls "performSubstitution" to actually perform the particle replacement. This method is called automatically by CParticleFilter::execute, andshould not be invoked manually normally. To just obtaining the sequence of resampled indexes from a sequence of weights, use "resample"

See also:: resample

virtual void mrpt::bayes::CParticleFilterCapable::performSubstitution ( const std::vector< size_t > & indx ) [pure virtual, inherited]

Performs the substitution for internal use of resample in particle filter algorithm, don't call it directly.

Parameters:

indx	The indices of current m_particles to be saved as the new m_particles set.

double mrpt::slam::CMultiMetricMapPDF::PF_SLAM_computeObservationLikelihoodForParticle	(	const CParticleFilter::TParticleFilterOptions &	PF_options,
		const size_t	particleIndexForMap,
		const CSensoryFrame &	observation,
		const CPose3D &	x
	)		const `[virtual]`

Evaluate the observation likelihood for one particle at a given location.

Implements mrpt::slam::PF_implementation< CRBPFParticleData, CMultiMetricMapPDF >.

virtual void mrpt::slam::PF_implementation::PF_SLAM_implementation_custom_update_particle_with_new_pose	(	PARTICLE_TYPE *	particleData,
		const TPose3D &	newPose
	)		const `[pure virtual, inherited]`

void mrpt::slam::CMultiMetricMapPDF::PF_SLAM_implementation_custom_update_particle_with_new_pose	(	CParticleDataContent *	particleData,
		const TPose3D &	newPose
	)		const

virtual bool mrpt::slam::PF_implementation::PF_SLAM_implementation_doWeHaveValidObservations	(	const typename CParticleFilterData< PARTICLE_TYPE >::CParticleList &	particles,
		const CSensoryFrame *	sf
	)		const `[inline, virtual, inherited]`

Definition at line 258 of file PF_implementations_data.h.

bool mrpt::slam::CMultiMetricMapPDF::PF_SLAM_implementation_doWeHaveValidObservations	(	const CParticleList &	particles,
		const CSensoryFrame *	sf
	)		const

bool mrpt::slam::PF_implementation::PF_SLAM_implementation_gatherActionsCheckBothActObs	(	const CActionCollection *	actions,
		const CSensoryFrame *	sf
	)		`[inherited]`

Auxiliary method called by PF implementations: return true if we have both action & observation, otherwise, return false AND accumulate the odometry so when we have an observation we didn't lose a thing.

On return=true, the "m_movementDrawer" member is loaded and ready to draw samples of the increment of pose since last step. This method is smart enough to accumulate CActionRobotMovement2D or CActionRobotMovement3D, whatever comes in.

void mrpt::slam::PF_implementation::PF_SLAM_implementation_pfAuxiliaryPFOptimal	(	const CActionCollection *	actions,
		const CSensoryFrame *	sf,
		const CParticleFilter::TParticleFilterOptions &	PF_options,
		const TKLDParams &	KLD_options
	)		`[protected, inherited]`

A generic implementation of the PF method "prediction_and_update_pfAuxiliaryPFOptimal" (optimal sampling with rejection sampling approximation), common to both localization and mapping.

BINTYPE: TPoseBin or whatever to discretize the sample space for KLD-sampling.

This method implements optimal sampling with a rejection sampling-based approximation of the true posterior. For details, see the papers:

J.-L. Blanco, J. González, and J.-A. Fernández-Madrigal, "An Optimal Filtering Algorithm for Non-Parametric Observation Models in Robot Localization," in Proc. IEEE International Conference on Robotics and Automation (ICRA'08), 2008, pp. 461–466.

BINTYPE: TPoseBin or whatever to discretize the sample space for KLD-sampling.

This method implements optimal sampling with a rejection sampling-based approximation of the true posterior. For details, see the papers:

J.-L. Blanco, J. GonzÃ¡lez, and J.-A. FernÃ¡ndez-Madrigal, "An Optimal Filtering Algorithm for Non-Parametric Observation Models in Robot Localization," in Proc. IEEE International Conference on Robotics and Automation (ICRA'08), 2008, pp. 461â€“466.

void mrpt::slam::PF_implementation::PF_SLAM_implementation_pfAuxiliaryPFStandard	(	const CActionCollection *	actions,
		const CSensoryFrame *	sf,
		const CParticleFilter::TParticleFilterOptions &	PF_options,
		const TKLDParams &	KLD_options
	)		`[protected, inherited]`

A generic implementation of the PF method "prediction_and_update_pfAuxiliaryPFStandard" (Auxiliary particle filter with the standard proposal), common to both localization and mapping.

BINTYPE: TPoseBin or whatever to discretize the sample space for KLD-sampling.

This method is described in the paper: Pitt, M.K.; Shephard, N. (1999). "Filtering Via Simulation: Auxiliary Particle Filters". Journal of the American Statistical Association 94 (446): 590–591. doi:10.2307/2670179.

BINTYPE: TPoseBin or whatever to discretize the sample space for KLD-sampling.

This method is described in the paper: Pitt, M.K.; Shephard, N. (1999). "Filtering Via Simulation: Auxiliary Particle Filters". Journal of the American Statistical Association 94 (446): 590â€“591. doi:10.2307/2670179.

void mrpt::slam::PF_implementation::PF_SLAM_implementation_pfStandardProposal	(	const CActionCollection *	actions,
		const CSensoryFrame *	sf,
		const CParticleFilter::TParticleFilterOptions &	PF_options,
		const TKLDParams &	KLD_options
	)		`[protected, inherited]`

A generic implementation of the PF method "pfStandardProposal" (standard proposal distribution, that is, a simple SIS particle filter), common to both localization and mapping.

BINTYPE: TPoseBin or whatever to discretize the sample space for KLD-sampling.

virtual void mrpt::slam::PF_implementation::PF_SLAM_implementation_replaceByNewParticleSet	(	typename CParticleFilterData< PARTICLE_TYPE >::CParticleList &	old_particles,
		const vector< TPose3D > &	newParticles,
		const vector_double &	newParticlesWeight,
		const vector< size_t > &	newParticlesDerivedFromIdx
	)		const `[inline, virtual, inherited]`

This is the default algorithm to efficiently replace one old set of samples by another new set.

The method uses pointers to make fast copies the first time each particle is duplicated, then makes real copies for the next ones.

Note that more efficient specializations might exist for specific particle data structs.

Definition at line 193 of file PF_implementations_data.h.

bool mrpt::slam::CMultiMetricMapPDF::PF_SLAM_implementation_skipRobotMovement ( ) const [virtual]

Make a specialization if needed, eg.

in the first step in SLAM.

Reimplemented from mrpt::slam::PF_implementation< CRBPFParticleData, CMultiMetricMapPDF >.

static double mrpt::slam::PF_implementation::PF_SLAM_particlesEvaluator_AuxPFOptimal	(	const CParticleFilter::TParticleFilterOptions &	PF_options,
		const CParticleFilterCapable *	obj,
		size_t	index,
		const void *	action,
		const void *	observation
	)		`[static, protected, inherited]`

static double mrpt::slam::PF_implementation::PF_SLAM_particlesEvaluator_AuxPFStandard	(	const CParticleFilter::TParticleFilterOptions &	PF_options,
		const CParticleFilterCapable *	obj,
		size_t	index,
		const void *	action,
		const void *	observation
	)		`[static, protected, inherited]`

Compute w[i]·p(z_t | mu_t^i), with mu_t^i being the mean of the new robot pose.

Compute w[i]Â·p(z_t | mu_t^i), with mu_t^i being the mean of the new robot pose.

Parameters:

action	MUST be a "const CPose3D*"
observation	MUST be a "const CSensoryFrame*"

void mrpt::bayes::CParticleFilterCapable::prediction_and_update	(	const mrpt::slam::CActionCollection *	action,
		const mrpt::slam::CSensoryFrame *	observation,
		const bayes::CParticleFilter::TParticleFilterOptions &	PF_options
	)		`[inherited]`

Performs the prediction stage of the Particle Filter.

This method simply selects the appropiate protected method according to the particle filter algorithm to run.

See also:: prediction_and_update_pfStandardProposal,prediction_and_update_pfAuxiliaryPFStandard,prediction_and_update_pfOptimalProposal,prediction_and_update_pfAuxiliaryPFOptimal

void mrpt::slam::CMultiMetricMapPDF::prediction_and_update_pfAuxiliaryPFOptimal	(	const mrpt::slam::CActionCollection *	action,
		const mrpt::slam::CSensoryFrame *	observation,
		const bayes::CParticleFilter::TParticleFilterOptions &	PF_options
	)		`[protected, virtual]`

The PF algorithm implementation.

Reimplemented from mrpt::bayes::CParticleFilterCapable.

virtual void mrpt::bayes::CParticleFilterCapable::prediction_and_update_pfAuxiliaryPFStandard	(	const mrpt::slam::CActionCollection *	action,
		const mrpt::slam::CSensoryFrame *	observation,
		const bayes::CParticleFilter::TParticleFilterOptions &	PF_options
	)		`[protected, virtual, inherited]`

Performs the particle filter prediction/update stages for the algorithm "pfAuxiliaryPFStandard" (if not implemented in heritated class, it will raise a 'non-implemented' exception).

See also:: prediction_and_update

Reimplemented in mrpt::slam::CMonteCarloLocalization2D, and mrpt::slam::CMonteCarloLocalization3D.

void mrpt::slam::CMultiMetricMapPDF::prediction_and_update_pfOptimalProposal	(	const mrpt::slam::CActionCollection *	action,
		const mrpt::slam::CSensoryFrame *	observation,
		const bayes::CParticleFilter::TParticleFilterOptions &	PF_options
	)		`[protected, virtual]`

The PF algorithm implementation.

Reimplemented from mrpt::bayes::CParticleFilterCapable.

void mrpt::slam::CMultiMetricMapPDF::prediction_and_update_pfStandardProposal	(	const mrpt::slam::CActionCollection *	action,
		const mrpt::slam::CSensoryFrame *	observation,
		const bayes::CParticleFilter::TParticleFilterOptions &	PF_options
	)		`[protected, virtual]`

The PF algorithm implementation.

Reimplemented from mrpt::bayes::CParticleFilterCapable.

void mrpt::bayes::CParticleFilterCapable::prepareFastDrawSample	(	const bayes::CParticleFilter::TParticleFilterOptions &	PF_options,
		TParticleProbabilityEvaluator	partEvaluator = `defaultEvaluator`,
		const void *	action = `NULL`,
		const void *	observation = `NULL`
	)		const `[inherited]`

Prepares data structures for calling fastDrawSample method next.

This method must be called once before using "fastDrawSample" (calling this more than once has no effect, but it takes time for nothing!) The behavior depends on the configuration of the PF (see CParticleFilter::TParticleFilterOptions):

DYNAMIC SAMPLE SIZE=NO: In this case this method fills out an internal array (m_fastDrawAuxiliary.alreadyDrawnIndexes) with the random indexes generated according to the selected resample scheme in TParticleFilterOptions. Those indexes are read sequentially by subsequent calls to fastDrawSample.
DYNAMIC SAMPLE SIZE=YES: Then:
- If TParticleFilterOptions.resamplingMethod = prMultinomial, the internal buffers will be filled out (m_fastDrawAuxiliary.CDF, CDF_indexes & PDF) and then fastDrawSample can be called an arbitrary number of times to generate random indexes.
- For the rest of resampling algorithms, an exception will be raised since they are not appropriate for a dynamic (unknown in advance) number of particles.

The function pointed by "partEvaluator" should take into account the particle filter algorithm selected in "m_PFAlgorithm". If called without arguments (defaultEvaluator), the default behavior is to draw samples with a probability proportional to their current weights. The action and the observation are declared as "void*" for a greater flexibility. For a more detailed information see the Particle Filter tutorial. Custom supplied "partEvaluator" functions must take into account the previous particle weight, i.e. multiplying the current observation likelihood by the weights.

See also:: fastDrawSample

Referenced by mrpt::slam::PF_implementation::PF_SLAM_implementation_pfAuxiliaryPFStandardAndOptimal().

virtual void mrpt::utils::CSerializable::readFromStream	(	mrpt::utils::CStream &	in,
		int	version
	)		`[protected, pure virtual, inherited]`

Introduces a pure virtual method responsible for loading from a CStream This can not be used directly be users, instead use "stream >> object;" for reading it from a stream or "stream >> object_ptr;" if the class is unknown apriori.

Parameters:

in	The input binary stream where the object data must read from.
version	The version of the object stored in the stream: use this version number in your code to know how to read the incoming data.

Exceptions:

std::exception On any error, see CStream::ReadBuffer

See also:: CStream

Implemented in mrpt::math::CMatrixD, and mrpt::math::CMatrix.

void mrpt::bayes::CParticleFilterData::readParticlesFromStream ( utils::CStream & in ) [inline, inherited]

Reads the sequence of particles and their weights from a stream (requires T implementing CSerializable).

See also:: writeParticlesToStream

Definition at line 107 of file CParticleFilterData.h.

References MRPT_START, mrpt::bayes::CParticleFilterData::clearParticles(), mrpt::bayes::CParticleFilterData::m_particles, and MRPT_END.

void mrpt::slam::CMultiMetricMapPDF::rebuildAverageMap ( ) [private]

Rebuild the "expected" grid map.

Used internally, do not call

void mrpt::slam::CMultiMetricMapPDF::saveCurrentPathEstimationToTextFile ( const std::string & fil )

A logging utility: saves the current path estimation for each particle in a text file (a row per particle, each 3-column-entry is a set [x,y,phi], respectively).

virtual void mrpt::bayes::CParticleFilterCapable::setW	(	size_t	i,
		double	w
	)		`[pure virtual, inherited]`

Modifies i'th particle (logarithm) weight, where first one is index 0.

void mrpt::slam::CMultiMetricMapPDF::updateSensoryFrameSequence ( )

Update the poses estimation of the member "SFs" according to the current path belief.

void mrpt::bayes::CParticleFilterData::writeParticlesToStream ( utils::CStream & out ) const [inline, inherited]

Dumps the sequence of particles and their weights to a stream (requires T implementing CSerializable).

See also:: readParticlesFromStream

Definition at line 93 of file CParticleFilterData.h.

References MRPT_START, mrpt::bayes::CParticleFilterData::m_particles, and MRPT_END.

virtual void mrpt::utils::CSerializable::writeToStream	(	mrpt::utils::CStream &	out,
		int *	getVersion
	)		const `[protected, pure virtual, inherited]`

Introduces a pure virtual method responsible for writing to a CStream.

This can not be used directly be users, instead use "stream << object;" for writing it to a stream.

Parameters:

out	The output binary stream where object must be dumped.
getVersion	If NULL, the object must be dumped. If not, only the version of the object dump must be returned in this pointer. This enables the versioning of objects dumping and backward compatibility with previously stored data.

Exceptions:

std::exception On any error, see CStream::WriteBuffer