MRPT  1.9.9
mrpt::bayes::CParticleFilterCapable Class Referenceabstract

Detailed Description

This virtual class defines the interface that any particles based PDF class must implement in order to be executed by a mrpt::bayes::CParticleFilter.

See the Particle Filter tutorial explaining how to use the particle filter-related classes.

See also
CParticleFilter, CParticleFilterData

Definition at line 31 of file CParticleFilterCapable.h.

#include <mrpt/bayes/CParticleFilterCapable.h>

Inheritance diagram for mrpt::bayes::CParticleFilterCapable:

Classes

struct  TFastDrawAuxVars
 Auxiliary vectors, see CParticleFilterCapable::prepareFastDrawSample for more information. More...
 

Public Types

using TParticleProbabilityEvaluator = double(*)(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". More...
 

Public Member Functions

 CParticleFilterCapable ()
 
virtual ~CParticleFilterCapable ()=default
 Virtual destructor. More...
 
void prepareFastDrawSample (const bayes::CParticleFilter::TParticleFilterOptions &PF_options, TParticleProbabilityEvaluator partEvaluator=defaultEvaluator, const void *action=nullptr, const void *observation=nullptr) const
 Prepares data structures for calling fastDrawSample method next. More...
 
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). More...
 
virtual double getW (size_t i) const =0
 Access to i'th particle (logarithm) weight, where first one is index 0. More...
 
virtual void setW (size_t i, double w)=0
 Modifies i'th particle (logarithm) weight, where first one is index 0. More...
 
virtual size_t particlesCount () const =0
 Get the m_particles count. More...
 
void prediction_and_update (const mrpt::obs::CActionCollection *action, const mrpt::obs::CSensoryFrame *observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
 Performs the prediction stage of the Particle Filter. More...
 
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. More...
 
virtual double normalizeWeights (double *out_max_log_w=nullptr)=0
 Normalize the (logarithmic) weights, such as the maximum weight is zero. More...
 
virtual double ESS () const =0
 Returns the normalized ESS (Estimated Sample Size), in the range [0,1]. More...
 
void performResampling (const bayes::CParticleFilter::TParticleFilterOptions &PF_options, size_t out_particle_count=0)
 Performs a resample of the m_particles, using the method selected in the constructor. More...
 

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. More...
 
static void computeResampling (CParticleFilter::TParticleResamplingAlgorithm method, const std::vector< double > &in_logWeights, std::vector< size_t > &out_indexes, size_t out_particle_count=0)
 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. More...
 
static void log2linearWeights (const std::vector< double > &in_logWeights, std::vector< double > &out_linWeights)
 A static method to compute the linear, normalized (the sum the unity) weights from log-weights. More...
 

Protected Member Functions

virtual void prediction_and_update_pfStandardProposal (const mrpt::obs::CActionCollection *action, const mrpt::obs::CSensoryFrame *observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
 Performs the particle filter prediction/update stages for the algorithm "pfStandardProposal" (if not implemented in heritated class, it will raise a 'non-implemented' exception). More...
 
virtual void prediction_and_update_pfAuxiliaryPFStandard (const mrpt::obs::CActionCollection *action, const mrpt::obs::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). More...
 
virtual void prediction_and_update_pfOptimalProposal (const mrpt::obs::CActionCollection *action, const mrpt::obs::CSensoryFrame *observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
 Performs the particle filter prediction/update stages for the algorithm "pfOptimalProposal" (if not implemented in heritated class, it will raise a 'non-implemented' exception). More...
 
virtual void prediction_and_update_pfAuxiliaryPFOptimal (const mrpt::obs::CActionCollection *action, const mrpt::obs::CSensoryFrame *observation, const bayes::CParticleFilter::TParticleFilterOptions &PF_options)
 Performs the particle filter prediction/update stages for the algorithm "pfAuxiliaryPFOptimal" (if not implemented in heritated class, it will raise a 'non-implemented' exception). More...
 

Protected Attributes

TFastDrawAuxVars m_fastDrawAuxiliary
 Auxiliary vectors, see CParticleFilterCapable::prepareFastDrawSample for more information. More...
 

Static Private Attributes

static const unsigned PARTICLE_FILTER_CAPABLE_FAST_DRAW_BINS
 

Friends

class CParticleFilter
 

Member Typedef Documentation

◆ TParticleProbabilityEvaluator

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
indexThis is the index of the particle its probability is being computed.
actionThe value of this is the parameter passed to "prepareFastDrawSample"
observationThe 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 60 of file CParticleFilterCapable.h.

Constructor & Destructor Documentation

◆ CParticleFilterCapable()

mrpt::bayes::CParticleFilterCapable::CParticleFilterCapable ( )
inline

Definition at line 39 of file CParticleFilterCapable.h.

◆ ~CParticleFilterCapable()

virtual mrpt::bayes::CParticleFilterCapable::~CParticleFilterCapable ( )
virtualdefault

Virtual destructor.

Member Function Documentation

◆ computeResampling()

void CParticleFilterCapable::computeResampling ( CParticleFilter::TParticleResamplingAlgorithm  method,
const std::vector< double > &  in_logWeights,
std::vector< size_t > &  out_indexes,
size_t  out_particle_count = 0 
)
static

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

Parameters
[in]out_particle_countThe desired number of output particles after resampling; 0 means don't modify the current number.
See also
performResampling

Definition at line 61 of file CParticleFilterCapable.cpp.

References ASSERT_, mrpt::random::CRandomGenerator::drawUniform(), mrpt::random::CRandomGenerator::drawUniformVector(), mrpt::format(), mrpt::random::getRandomGenerator(), mrpt::math::maximum(), MRPT_END, MRPT_START, mrpt::bayes::CParticleFilter::prMultinomial, mrpt::bayes::CParticleFilter::prResidual, mrpt::bayes::CParticleFilter::prStratified, mrpt::bayes::CParticleFilter::prSystematic, R, and THROW_EXCEPTION.

Here is the call graph for this function:

◆ defaultEvaluator()

static double mrpt::bayes::CParticleFilterCapable::defaultEvaluator ( const bayes::CParticleFilter::TParticleFilterOptions PF_options,
const CParticleFilterCapable obj,
size_t  index,
const void action,
const void observation 
)
inlinestatic

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 68 of file CParticleFilterCapable.h.

References MRPT_UNUSED_PARAM.

◆ ESS()

◆ fastDrawSample()

size_t CParticleFilterCapable::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).

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

Definition at line 486 of file CParticleFilterCapable.cpp.

References mrpt::bayes::CParticleFilter::TParticleFilterOptions::adaptiveSampleSize, mrpt::random::CRandomGenerator::drawUniform(), mrpt::random::getRandomGenerator(), MRPT_END, MRPT_END_WITH_CLEAN_UP, MRPT_START, mrpt::bayes::CParticleFilter::prMultinomial, mrpt::bayes::CParticleFilter::TParticleFilterOptions::resamplingMethod, and THROW_EXCEPTION.

Referenced by mrpt::hmtslam::CLSLAM_RBPF_2DLASER::prediction_and_update_pfAuxiliaryPFOptimal().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ getW()

◆ log2linearWeights()

void CParticleFilterCapable::log2linearWeights ( const std::vector< double > &  in_logWeights,
std::vector< double > &  out_linWeights 
)
static

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

See also
performResampling

Definition at line 552 of file CParticleFilterCapable.cpp.

References ASSERT_, MRPT_END, and MRPT_START.

◆ normalizeWeights()

◆ particlesCount()

◆ performResampling()

void CParticleFilterCapable::performResampling ( const bayes::CParticleFilter::TParticleFilterOptions PF_options,
size_t  out_particle_count = 0 
)

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"

Parameters
[in]out_particle_countThe desired number of output particles after resampling; 0 means don't modify the current number.
See also
resample

Definition at line 29 of file CParticleFilterCapable.cpp.

References ASSERT_, MRPT_END, MRPT_START, and mrpt::bayes::CParticleFilter::TParticleFilterOptions::resamplingMethod.

◆ performSubstitution()

◆ prediction_and_update()

◆ prediction_and_update_pfAuxiliaryPFOptimal()

void CParticleFilterCapable::prediction_and_update_pfAuxiliaryPFOptimal ( const mrpt::obs::CActionCollection action,
const mrpt::obs::CSensoryFrame observation,
const bayes::CParticleFilter::TParticleFilterOptions PF_options 
)
protectedvirtual

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

See also
prediction_and_update

Reimplemented in mrpt::hmtslam::CLocalMetricHypothesis, mrpt::slam::CMonteCarloLocalization2D, mrpt::slam::CMonteCarloLocalization3D, and mrpt::maps::CMultiMetricMapPDF.

Definition at line 351 of file CParticleFilterCapable.cpp.

References MRPT_UNUSED_PARAM, and THROW_EXCEPTION.

◆ prediction_and_update_pfAuxiliaryPFStandard()

void CParticleFilterCapable::prediction_and_update_pfAuxiliaryPFStandard ( const mrpt::obs::CActionCollection action,
const mrpt::obs::CSensoryFrame observation,
const bayes::CParticleFilter::TParticleFilterOptions PF_options 
)
protectedvirtual

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, mrpt::maps::CMultiMetricMapPDF, and mrpt::slam::CMonteCarloLocalization3D.

Definition at line 322 of file CParticleFilterCapable.cpp.

References MRPT_UNUSED_PARAM, and THROW_EXCEPTION.

◆ prediction_and_update_pfOptimalProposal()

void CParticleFilterCapable::prediction_and_update_pfOptimalProposal ( const mrpt::obs::CActionCollection action,
const mrpt::obs::CSensoryFrame observation,
const bayes::CParticleFilter::TParticleFilterOptions PF_options 
)
protectedvirtual

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

See also
prediction_and_update

Reimplemented in mrpt::hmtslam::CLocalMetricHypothesis, and mrpt::maps::CMultiMetricMapPDF.

Definition at line 337 of file CParticleFilterCapable.cpp.

References MRPT_UNUSED_PARAM, and THROW_EXCEPTION.

◆ prediction_and_update_pfStandardProposal()

void CParticleFilterCapable::prediction_and_update_pfStandardProposal ( const mrpt::obs::CActionCollection action,
const mrpt::obs::CSensoryFrame observation,
const bayes::CParticleFilter::TParticleFilterOptions PF_options 
)
protectedvirtual

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

See also
prediction_and_update

Reimplemented in CRangeBearingParticleFilter, mrpt::slam::CMonteCarloLocalization2D, mrpt::maps::CMultiMetricMapPDF, and mrpt::slam::CMonteCarloLocalization3D.

Definition at line 307 of file CParticleFilterCapable.cpp.

References MRPT_UNUSED_PARAM, and THROW_EXCEPTION.

◆ prepareFastDrawSample()

void CParticleFilterCapable::prepareFastDrawSample ( const bayes::CParticleFilter::TParticleFilterOptions PF_options,
TParticleProbabilityEvaluator  partEvaluator = defaultEvaluator,
const void action = nullptr,
const void observation = nullptr 
) const

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

Definition at line 367 of file CParticleFilterCapable.cpp.

References mrpt::bayes::CParticleFilter::TParticleFilterOptions::adaptiveSampleSize, ASSERT_, mrpt::math::maximum(), MRPT_CHECK_NORMAL_NUMBER, MRPT_END, MRPT_END_WITH_CLEAN_UP, MRPT_START, mrpt::bayes::CParticleFilter::prMultinomial, mrpt::bayes::CParticleFilter::TParticleFilterOptions::resamplingMethod, and THROW_EXCEPTION.

Referenced by mrpt::hmtslam::CLSLAM_RBPF_2DLASER::prediction_and_update_pfAuxiliaryPFOptimal().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ setW()

Friends And Related Function Documentation

◆ CParticleFilter

friend class CParticleFilter
friend

Definition at line 33 of file CParticleFilterCapable.h.

Member Data Documentation

◆ m_fastDrawAuxiliary

TFastDrawAuxVars mrpt::bayes::CParticleFilterCapable::m_fastDrawAuxiliary
mutableprotected

Auxiliary vectors, see CParticleFilterCapable::prepareFastDrawSample for more information.

Definition at line 283 of file CParticleFilterCapable.h.

◆ PARTICLE_FILTER_CAPABLE_FAST_DRAW_BINS

const unsigned CParticleFilterCapable::PARTICLE_FILTER_CAPABLE_FAST_DRAW_BINS
staticprivate
Initial value:
=
20

Definition at line 36 of file CParticleFilterCapable.h.




Page generated by Doxygen 1.8.14 for MRPT 1.9.9 Git: 928d0abbd Sun Oct 13 02:28:49 2019 +0200 at dom oct 13 02:30:11 CEST 2019