Standard linear Kalman Filter. More...

#include <shark/Models/KalmanFilter.h>

Inheritance diagram for shark::KalmanFilter:

Collaboration diagram for shark::KalmanFilter:

Public Member Functions
	KalmanFilter (size_t modelSize)
	Constructor. More...

std::string	name () const
	From INameable: return the class name. More...

RealVector	parameterVector () const
	Return the parameter vector. More...

void	setParameterVector (RealVector const &newParameters)
	Set the parameter vector. More...

void	eval (const RealVector &pattern, RealVector &output)

RealMatrix &	stateDynamics ()
	returns the state transition matrix 'F' More...

const RealMatrix &	stateDynamics () const

RealMatrix &	stateNoise ()
	returns the state noise covariance matrix 'Q' More...

const RealMatrix &	stateNoise () const

RealMatrix &	observationModel ()
	sets the observation matrix 'H' More...

const RealMatrix &	observationModel () const

RealMatrix &	observationNoise ()
	sets the observation noise covriance matrix 'R' More...

const RealMatrix &	observationNoise () const

void	setStateVector (const RealVector &stateVec)
	sets the initial state vector More...

void	setStateErrorCov (const RealMatrix &stateVec)
	sets the initial covariance matrix of state estimation error More...

void	predict ()
	next time step / forward propgation (calculation of a-priori estimates) More...

void	updateEstimate (const RealVector &z)
	update state vector; parameter z: measurent vector More...

RealVector &	statePrediction ()
	returs a-priori (predicted) state estimate (updated in 'propagate()') More...

const RealVector &	statePrediction () const

RealMatrix &	predictedCovariance ()
	returns a-priori (predicted) covariance of state estimation error More...

const RealMatrix &	predictedCovariance () const

RealMatrix &	observedCovariance ()
	returns a-posteriori covariance of state estimation error (after z(t) has been observed) More...

const RealVector &	predictionError () const
	returns error in predicted measurements More...

Public Member Functions inherited from shark::AbstractModel< RealVector, RealVector >
	AbstractModel ()

virtual	~AbstractModel ()

const Features &	features () const

virtual void	updateFeatures ()

bool	hasFirstParameterDerivative () const
	Returns true when the first parameter derivative is implemented. More...

bool	hasSecondParameterDerivative () const
	Returns true when the second parameter derivative is implemented. More...

bool	hasFirstInputDerivative () const
	Returns true when the first input derivative is implemented. More...

bool	hasSecondInputDerivative () const
	Returns true when the second parameter derivative is implemented. More...

bool	isSequential () const

virtual boost::shared_ptr< State >	createState () const
	Creates an internal state of the model. More...

virtual void	read (InArchive &archive)
	From ISerializable, reads a model from an archive. More...

virtual void	write (OutArchive &archive) const
	writes a model to an archive More...

virtual void	eval (BatchInputType const &patterns, BatchOutputType &outputs) const
	Standard interface for evaluating the response of the model to a batch of patterns. More...

virtual void	eval (BatchInputType const &patterns, BatchOutputType &outputs, State &state) const =0
	Standard interface for evaluating the response of the model to a batch of patterns. More...

virtual void	eval (InputType const &pattern, OutputType &output) const
	Standard interface for evaluating the response of the model to a single pattern. More...

Data< OutputType >	operator() (Data< InputType > const &patterns) const
	Model evaluation as an operator for a whole dataset. This is a convenience function. More...

OutputType	operator() (InputType const &pattern) const
	Model evaluation as an operator for a single pattern. This is a convenience function. More...

BatchOutputType	operator() (BatchInputType const &patterns) const
	Model evaluation as an operator for a single pattern. This is a convenience function. More...

virtual void	weightedParameterDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, State const &state, RealVector &derivative) const
	calculates the weighted sum of derivatives w.r.t the parameters. More...

virtual void	weightedParameterDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, Batch< RealMatrix >::type const &errorHessian, State const &state, RealVector &derivative, RealMatrix &hessian) const
	calculates the weighted sum of derivatives w.r.t the parameters More...

virtual void	weightedInputDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, State const &state, BatchInputType &derivative) const
	calculates the weighted sum of derivatives w.r.t the inputs More...

virtual void	weightedInputDerivative (BatchInputType const &pattern, BatchOutputType const &coefficients, typename Batch< RealMatrix >::type const &errorHessian, State const &state, RealMatrix &derivative, Batch< RealMatrix >::type &hessian) const
	calculates the weighted sum of derivatives w.r.t the inputs More...

virtual void	weightedDerivatives (BatchInputType const &patterns, BatchOutputType const &coefficients, State const &state, RealVector &parameterDerivative, BatchInputType &inputDerivative) const
	calculates weighted input and parameter derivative at the same time More...

Public Member Functions inherited from shark::IParameterizable
virtual	~IParameterizable ()

virtual std::size_t	numberOfParameters () const
	Return the number of parameters. More...

Public Member Functions inherited from shark::IConfigurable
virtual	~IConfigurable ()
	Virtual d'tor. More...

virtual void	configure (const PropertyTree &node)
	Configures the component given a property tree. More...

Public Member Functions inherited from shark::INameable
virtual	~INameable ()

Public Member Functions inherited from shark::ISerializable
virtual	~ISerializable ()
	Virtual d'tor. More...

void	load (InArchive &archive, unsigned int version)
	Versioned loading of components, calls read(...). More...

void	save (OutArchive &archive, unsigned int version) const
	Versioned storing of components, calls write(...). More...

	BOOST_SERIALIZATION_SPLIT_MEMBER ()

Protected Attributes
RealMatrix	m_stateTransition
	state transition matrix More...

RealMatrix	m_stateNoiseCov
	state noise covariance matrix More...

RealMatrix	m_observationModel
	observation matrix More...

RealMatrix	m_observationNoise
	observation noise covariance matrix More...

RealVector	m_predictedState
	a-priori (predicted) state estimate at time t (before z(t) has been observed) More...

RealVector	m_state
	a-posteriori state at timte t (after observation of z(t)) More...

RealMatrix	m_predictedCovariance
	a-priori (predicted) covariance of state estimation error More...

RealMatrix	m_observedCovariance
	a-posteriori covariance of state estimation error (after z(t) has been observed) More...

RealVector	m_predictionError
	error in predicted measurements More...

size_t	m_inputSize

Protected Attributes inherited from shark::AbstractModel< RealVector, RealVector >
Features	m_features

Additional Inherited Members
Public Types inherited from shark::AbstractModel< RealVector, RealVector >
enum	Feature

typedef RealVector	InputType
	Defines the input type of the model. More...

typedef RealVector	OutputType
	Defines the output type of the model. More...

typedef Batch< InputType >::type	BatchInputType
	defines the batch type of the input type. More...

typedef Batch< OutputType >::type	BatchOutputType
	defines the batch type of the output type More...

typedef TypedFlags< Feature >	Features

typedef TypedFeatureNotAvailableException < Feature >	FeatureNotAvailableException

Detailed Description

Standard linear Kalman Filter.

The implementation aims at providing flexibility w.r.t. process and measurement model; no restrictive assumptions concerning specific properties of the matrices are made (although in practical applications often noise covariances have diagonal form). The implementation is not optimized at all for computational efficiency.

For making the algorithm robust against numerical inaccuraries, the update equation for the state error covariance matrix 'P' given in "Applied Optimal Estimation; The Analytic Sciences Corporation; The M.I.T. Press, pp. 305" is used. For very complex modells a square root form (e.g. Cholesky) based implementation might be more appropriate due to its increased numerical robustness.

Process Modell:
------------—
State Dynamics: \( x_t = F*x_{t-1} + n_x; n_x ~ N(0,Q) \)
Observation : \( z_t = H*x_t + n_z; n_z ~ N(0,R) \)

Before filtering the model matrices F,Q,H,R must be set. Furthermore the initial filter states, i.e. the initial state vector x(0) and the initial covariance matrix of state estimation error P(0) must be set.

As a Model subclass the KalmanFilter holds its current a-posteriori state estimate at time t (after z(t) has been observed) in the Model::parameter vector.

Definition at line 38 of file KalmanFilter.h.

Constructor & Destructor Documentation

KalmanFilter::KalmanFilter ( size_t modelSize )

Constructor.

Definition at line 42 of file KalmanFilter.cpp.

References m_inputSize.

Member Function Documentation

void KalmanFilter::eval	(	const RealVector &	pattern,
		RealVector &	output
	)

The model first runs the predict phase. Then it updates its internal estimates using the input. Finally the a-posteriori state estimate, that is, the model parameter vector is returned as output.

Definition at line 46 of file KalmanFilter.cpp.

References m_state, predict(), and updateEstimate().

std::string shark::KalmanFilter::name ( ) const

inlinevirtual

From INameable: return the class name.

Reimplemented from shark::INameable.

Definition at line 45 of file KalmanFilter.h.

RealMatrix& shark::KalmanFilter::observationModel ( )

inline

sets the observation matrix 'H'

Definition at line 90 of file KalmanFilter.h.

References m_observationModel.

const RealMatrix& shark::KalmanFilter::observationModel ( ) const

inline

Definition at line 93 of file KalmanFilter.h.

References m_observationModel.

RealMatrix& shark::KalmanFilter::observationNoise ( )

inline

sets the observation noise covriance matrix 'R'

Definition at line 99 of file KalmanFilter.h.

References m_observationNoise.

const RealMatrix& shark::KalmanFilter::observationNoise ( ) const

inline

Definition at line 102 of file KalmanFilter.h.

References m_observationNoise.

RealMatrix& shark::KalmanFilter::observedCovariance ( )

inline

returns a-posteriori covariance of state estimation error (after z(t) has been observed)

Definition at line 152 of file KalmanFilter.h.

References m_observedCovariance.

RealVector shark::KalmanFilter::parameterVector ( ) const

inlinevirtual

Return the parameter vector.

Reimplemented from shark::IParameterizable.

Definition at line 48 of file KalmanFilter.h.

void KalmanFilter::predict ( )

next time step / forward propgation (calculation of a-priori estimates)

Definition at line 61 of file KalmanFilter.cpp.

References m_observedCovariance, m_predictedCovariance, m_predictedState, m_state, m_stateNoiseCov, m_stateTransition, shark::blas::prod(), and shark::blas::trans().

Referenced by eval().

RealMatrix& shark::KalmanFilter::predictedCovariance ( )

inline

returns a-priori (predicted) covariance of state estimation error

Definition at line 144 of file KalmanFilter.h.

References m_predictedCovariance.

const RealMatrix& shark::KalmanFilter::predictedCovariance ( ) const

inline

Definition at line 147 of file KalmanFilter.h.

References m_predictedCovariance.

const RealVector& shark::KalmanFilter::predictionError ( ) const

inline

returns error in predicted measurements

Definition at line 158 of file KalmanFilter.h.

References m_predictionError.

void shark::KalmanFilter::setParameterVector ( RealVector const & newParameters )

inlinevirtual

Set the parameter vector.

Reimplemented from shark::IParameterizable.

Definition at line 51 of file KalmanFilter.h.

void KalmanFilter::setStateErrorCov ( const RealMatrix & stateVec )

sets the initial covariance matrix of state estimation error

Definition at line 56 of file KalmanFilter.cpp.

References m_observedCovariance, and m_predictedCovariance.

void KalmanFilter::setStateVector ( const RealVector & stateVec )

sets the initial state vector

Definition at line 51 of file KalmanFilter.cpp.

References m_predictedState, and m_state.

RealMatrix& shark::KalmanFilter::stateDynamics ( )

inline

returns the state transition matrix 'F'

Definition at line 69 of file KalmanFilter.h.

References m_stateTransition.

const RealMatrix& shark::KalmanFilter::stateDynamics ( ) const

inline

Definition at line 72 of file KalmanFilter.h.

References m_stateTransition.

RealMatrix& shark::KalmanFilter::stateNoise ( )

inline

returns the state noise covariance matrix 'Q'

Definition at line 77 of file KalmanFilter.h.

References m_stateNoiseCov.

const RealMatrix& shark::KalmanFilter::stateNoise ( ) const

inline

Definition at line 80 of file KalmanFilter.h.

References m_stateNoiseCov.

RealVector& shark::KalmanFilter::statePrediction ( )

inline

returs a-priori (predicted) state estimate (updated in 'propagate()')

Definition at line 136 of file KalmanFilter.h.

References m_predictedState.

const RealVector& shark::KalmanFilter::statePrediction ( ) const

inline

Definition at line 139 of file KalmanFilter.h.

References m_predictedState.

void KalmanFilter::updateEstimate ( const RealVector & z )

update state vector; parameter z: measurent vector

Definition at line 72 of file KalmanFilter.cpp.

References shark::invert(), m_inputSize, m_observationModel, m_observationNoise, m_observedCovariance, m_predictedCovariance, m_predictedState, m_predictionError, m_state, shark::blas::prod(), and shark::blas::trans().

Referenced by eval().

Member Data Documentation

size_t shark::KalmanFilter::m_inputSize

protected

Definition at line 190 of file KalmanFilter.h.

Referenced by KalmanFilter(), and updateEstimate().

RealMatrix shark::KalmanFilter::m_observationModel

protected

observation matrix

Definition at line 171 of file KalmanFilter.h.

Referenced by observationModel(), and updateEstimate().

RealMatrix shark::KalmanFilter::m_observationNoise

protected

observation noise covariance matrix

Definition at line 174 of file KalmanFilter.h.

Referenced by observationNoise(), and updateEstimate().

RealMatrix shark::KalmanFilter::m_observedCovariance

protected

a-posteriori covariance of state estimation error (after z(t) has been observed)

Definition at line 185 of file KalmanFilter.h.

Referenced by observedCovariance(), predict(), setStateErrorCov(), and updateEstimate().

RealMatrix shark::KalmanFilter::m_predictedCovariance

protected

a-priori (predicted) covariance of state estimation error

Definition at line 182 of file KalmanFilter.h.

Referenced by predict(), predictedCovariance(), setStateErrorCov(), and updateEstimate().

RealVector shark::KalmanFilter::m_predictedState

protected

a-priori (predicted) state estimate at time t (before z(t) has been observed)

Definition at line 177 of file KalmanFilter.h.

Referenced by predict(), setStateVector(), statePrediction(), and updateEstimate().

RealVector shark::KalmanFilter::m_predictionError

protected

error in predicted measurements

Definition at line 188 of file KalmanFilter.h.

Referenced by predictionError(), and updateEstimate().

RealVector shark::KalmanFilter::m_state

protected

a-posteriori state at timte t (after observation of z(t))

Definition at line 179 of file KalmanFilter.h.

Referenced by eval(), predict(), setStateVector(), and updateEstimate().

RealMatrix shark::KalmanFilter::m_stateNoiseCov

protected

state noise covariance matrix

Definition at line 168 of file KalmanFilter.h.

Referenced by predict(), and stateNoise().

RealMatrix shark::KalmanFilter::m_stateTransition

protected

state transition matrix

Definition at line 165 of file KalmanFilter.h.

Referenced by predict(), and stateDynamics().

The documentation for this class was generated from the following files:

include/shark/Models/KalmanFilter.h
src/Models/KalmanFilter.cpp

Public Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation