Implementation of the Kalman filter method. More...

#include <KalmanFilter.h>

Inheritance diagram for nnp::KalmanFilter:

Collaboration diagram for nnp::KalmanFilter:

[legend]

Public Types
enum	KalmanType { KT_STANDARD , KT_FADINGMEMORY }
	Enumerate different Kalman filter types. More...

Public Member Functions
	KalmanFilter (std::size_t const sizeState, KalmanType const type)
	Kalman filter class constructor. More...

virtual	~KalmanFilter ()
	Destructor. More...

void	setSizeObservation (std::size_t const sizeObservation)
	Set observation vector size. More...

void	setState (double *state)
	Set pointer to current state. More...

void	setError (double const *const error)
	Set pointer to current error vector. More...

void	setError (double const *const error, std::size_t const size)
	Set pointer to current error vector. More...

void	setJacobian (double const *const jacobian)
	Set pointer to current Jacobi matrix. More...

void	setJacobian (double const *const jacobian, std::size_t const columns)
	Set pointer to current Jacobi matrix. More...

void	update ()
	Update error covariance matrix and state vector. More...

void	update (std::size_t const sizeObservation)
	Update error covariance matrix and state vector. More...

void	setParametersStandard (double const epsilon, double const q0, double const qtau, double const qmin, double const eta0, double const etatau, double const etamax)
	Set parameters for standard Kalman filter. More...

void	setParametersFadingMemory (double const epsilon, double const q0, double const qtau, double const qmin, double const lambda, double const nu)
	Set parameters for fading memory Kalman filter. More...

std::string	status (std::size_t epoch) const
	Status report. More...

std::vector< std::string >	statusHeader () const
	Header for status report file. More...

std::vector< std::string >	info () const
	Information about Kalman filter settings. More...

KalmanType	getType () const
	Getter for type. More...

std::size_t	getSizeObservation () const
	Getter for sizeObservation. More...

std::size_t	getNumUpdates () const
	Getter for numUpdates. More...

double	getEta () const
	Getter for eta. More...

double	getEpsilon () const
	Getter for epsilon. More...

double	getQ0 () const
	Getter for q0. More...

double	getQtau () const
	Getter for qtau. More...

double	getQmin () const
	Getter for qmin. More...

double	getLambda () const
	Getter for lambda. More...

double	getNu () const
	Getter for nu. More...

double	getGamma () const
	Getter for gamma. More...

Public Member Functions inherited from nnp::Updater
virtual void	setState (double *state)=0
	Set pointer to current state. More...

virtual void	setError (double const *const error, std::size_t const size=1)=0
	Set pointer to current error vector. More...

virtual void	setJacobian (double const *const jacobian, std::size_t const columns=1)=0
	Set pointer to current Jacobi matrix. More...

virtual void	update ()=0
	Perform single update of state vector. More...

virtual std::string	status (std::size_t epoch) const =0
	Status report. More...

virtual std::vector< std::string >	statusHeader () const =0
	Header for status report file. More...

virtual std::vector< std::string >	info () const =0
	Information about this updater. More...

virtual void	setupTiming (std::string const &prefix="upd")
	Activate detailed timing. More...

virtual void	resetTimingLoop ()
	Start a new timing loop (e.g. More...

virtual std::map< std::string, Stopwatch >	getTiming () const
	Return timings gathered in stopwatch map. More...

Private Attributes
KalmanType	type
	Kalman filter type. More...

std::size_t	sizeObservation
	Size of observation (measurement) vector. More...

std::size_t	numUpdates
	Total number of updates performed. More...

double	epsilon
	Error covariance initialization parameter \(\epsilon\). More...

double	q
	Process noise \(q\). More...

double	q0
	Process noise initial value \(q_0\). More...

double	qtau
	Process noise exponential decay parameter \(q_{\tau}\). More...

double	qmin
	Process noise minimum value \(q_{\text{min}}\). More...

double	eta
	Learning rate \(\eta\). More...

double	eta0
	Learning rate initial value \(\eta_0\). More...

double	etatau
	Learning rate exponential increase parameter \(\eta_{\tau}\). More...

double	etamax
	Learning rate maximum value \(\eta_{\text{max}}\). More...

double	lambda
	Forgetting factor for fading memory Kalman filter. More...

double	nu
	Parameter for fading memory Kalman filter. More...

double	gamma
	Forgetting gain factor gamma for fading memory Kalman filter. More...

Eigen::Map< Eigen::VectorXd > *	w
	State vector. More...

Eigen::Map< Eigen::VectorXd const > *	xi
	Error vector. More...

Eigen::Map< Eigen::MatrixXd const > *	H
	Derivative matrix. More...

Eigen::MatrixXd	P
	Error covariance matrix. More...

Eigen::MatrixXd	K
	Kalman gain matrix. More...

Eigen::MatrixXd	X
	Intermediate result X = P . H. More...

Additional Inherited Members
Protected Member Functions inherited from nnp::Updater
	Updater (std::size_t const sizeState)
	Constructor. More...

Protected Attributes inherited from nnp::Updater
bool	timing
	Whether detailed timing is enabled. More...

bool	timingReset
	Internal loop timer reset switch. More...

std::size_t	sizeState
	Number of neural network connections (weights + biases). More...

std::string	prefix
	Prefix for timing stopwatches. More...

std::map< std::string, Stopwatch >	sw
	Stopwatch map for timing. More...

Detailed Description

Implementation of the Kalman filter method.

Definition at line 31 of file KalmanFilter.h.

Member Enumeration Documentation

◆ KalmanType

enum nnp::KalmanFilter::KalmanType

Enumerate different Kalman filter types.

Enumerator
KT_STANDARD	Regular Kalman filter.
KT_FADINGMEMORY	Kalman filtering with fading memory modification.

Definition at line 35 of file KalmanFilter.h.

    {
        KT_STANDARD,
        KT_FADINGMEMORY
    };

Constructor & Destructor Documentation

◆ KalmanFilter()

KalmanFilter::KalmanFilter	(	std::size_t const	sizeState,
		KalmanType const	type
	)

Kalman filter class constructor.

Parameters

[in]	sizeState	Size of the state vector w.
[in]	type	Kalman filter type (see KalmanType).

Definition at line 27 of file KalmanFilter.cpp.

                                                  :
    Updater(sizeState),
    sizeObservation(0   ),
    numUpdates     (0   ),
    epsilon        (0.0 ),
    q              (0.0 ),
    q0             (0.0 ),
    qtau           (0.0 ),
    qmin           (0.0 ),
    eta            (0.0 ),
    eta0           (0.0 ),
    etatau         (0.0 ),
    etamax         (0.0 ),
    lambda         (0.0 ),
    nu             (0.0 ),
    gamma          (0.0 ),
    w              (NULL),
    xi             (NULL),
    H              (NULL)
{
    if (!(type == KT_STANDARD ||
          type == KT_FADINGMEMORY))
    {
        throw runtime_error("ERROR: Unknown Kalman filter type.\n");
    }
 
    if (sizeState < 1)
    {
        throw runtime_error("ERROR: Wrong Kalman filter dimensions.\n");
    }
 
    this->type            = type;
    sizeObservation = 1;
 
    w  = new Map<VectorXd      >(0, sizeState);
    xi = new Map<VectorXd const>(0, sizeObservation);
    H  = new Map<MatrixXd const>(0, sizeState, sizeObservation);
    P.resize(sizeState, sizeState);
    P.setIdentity();
    // Prevent problems with unallocated K when log starts.
    K.resize(sizeState, sizeObservation);
    K.setZero();
}

References H, K, KT_FADINGMEMORY, KT_STANDARD, P, sizeObservation, nnp::Updater::sizeState, type, w, and xi.

◆ ~KalmanFilter()

KalmanFilter::~KalmanFilter ( )

virtual

Destructor.

Definition at line 72 of file KalmanFilter.cpp.

73{

74}

Member Function Documentation

◆ setSizeObservation()

void KalmanFilter::setSizeObservation ( std::size_t const sizeObservation )

Set observation vector size.

Parameters

[in] size Size of the observation vector.

If the size of the observation vector is known in advance it can be set here.

Definition at line 76 of file KalmanFilter.cpp.

{
    sizeObservation = size;
 
    return;
}

References sizeObservation.

Referenced by nnp::Training::update().

Here is the caller graph for this function:

◆ setState()

void KalmanFilter::setState ( double * state )

virtual

Set pointer to current state.

Parameters

[in,out] state Pointer to state vector (weights vector), will be changed in-place upon calling update().

Implements nnp::Updater.

Definition at line 83 of file KalmanFilter.cpp.

{
    new (w) Map<VectorXd>(state, sizeState);
 
    return;
}

References nnp::Updater::sizeState, and w.

◆ setError() [1/2]

void KalmanFilter::setError ( double const *const error )

Set pointer to current error vector.

Parameters

[in] error Pointer to error (difference between reference and neural network potential output).

Definition at line 90 of file KalmanFilter.cpp.

{
    setError(error, sizeObservation);
 
    return;
}

References setError(), and sizeObservation.

Referenced by setError().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ setError() [2/2]

void nnp::KalmanFilter::setError	(	double const *const	error,
		std::size_t const	size
	)

virtual

Set pointer to current error vector.

Parameters

[in]	error	Pointer to error (difference between reference and neural network potential output).
[in]	size	Number of error vector entries.

Implements nnp::Updater.

◆ setJacobian() [1/2]

void KalmanFilter::setJacobian ( double const *const jacobian )

Set pointer to current Jacobi matrix.

Parameters

[in] jacobian Derivatives of error with respect to weights.

Definition at line 104 of file KalmanFilter.cpp.

{
    setJacobian(jacobian, sizeObservation);
 
    return;
}

References setJacobian(), and sizeObservation.

Referenced by setJacobian().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ setJacobian() [2/2]

void nnp::KalmanFilter::setJacobian	(	double const *const	jacobian,
		std::size_t const	columns
	)

virtual

Set pointer to current Jacobi matrix.

Parameters

[in]	jacobian	Derivatives of error with respect to weights.
[in]	columns	Number of gradients provided.

Note: If there are \(m\) errors and \(n\) weights, the Jacobi matrix is a \(n \times m\) matrix stored in column-major order.

Implements nnp::Updater.

◆ update() [1/2]

void KalmanFilter::update ( )

virtual

Update error covariance matrix and state vector.

Implements nnp::Updater.

Definition at line 119 of file KalmanFilter.cpp.

{
    update(sizeObservation);
 
    return;
}

References sizeObservation, and update().

Referenced by update().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ update() [2/2]

void nnp::KalmanFilter::update ( std::size_t const sizeObservation )

Update error covariance matrix and state vector.

Parameters

[in] sizeObservation Size of the observation vector.

◆ setParametersStandard()

void KalmanFilter::setParametersStandard	(	double const	epsilon,
		double const	q0,
		double const	qtau,
		double const	qmin,
		double const	eta0,
		double const	etatau,
		double const	etamax
	)

Set parameters for standard Kalman filter.

Parameters

[in]	epsilon	Error covariance initialization parameter \(\epsilon\).
[in]	q0	Process noise initial value \(q_0\).
[in]	qtau	Process noise exponential decay parameter \(q_{\tau}\).
[in]	qmin	Process noise minimum value \(q_{\text{min}}\).
[in]	eta0	Initial learning rate \(\eta_0\).
[in]	etatau	Learning rate exponential increase parameter \(\eta_{\tau}\).
[in]	etamax	Learning rate maximum value \(\eta_{\text{max}}\).

Definition at line 199 of file KalmanFilter.cpp.

{
    this->epsilon = epsilon;
    this->q0      = q0     ;
    this->qtau    = qtau   ;
    this->qmin    = qmin   ;
    this->eta0    = eta0   ;
    this->etatau  = etatau ;
    this->etamax  = etamax ;
 
    q = q0;
    eta = eta0;
    P /= epsilon;
 
    return;
}

References epsilon, eta, eta0, etamax, etatau, P, q, q0, qmin, and qtau.

Referenced by nnp::Training::setupTraining().

Here is the caller graph for this function:

◆ setParametersFadingMemory()

void KalmanFilter::setParametersFadingMemory	(	double const	epsilon,
		double const	q0,
		double const	qtau,
		double const	qmin,
		double const	lambda,
		double const	nu
	)

Set parameters for fading memory Kalman filter.

Parameters

[in]	epsilon	Error covariance initialization parameter \(\epsilon\).
[in]	q0	Process noise initial value \(q_0\).
[in]	qtau	Process noise exponential decay parameter \(q_{\tau}\).
[in]	qmin	Process noise minimum value \(q_{\text{min}}\).
[in]	lambda	Forgetting factor \(\lambda\).
[in]	nu	Fading memory parameter \(\nu\).

Definition at line 222 of file KalmanFilter.cpp.

{
    this->epsilon = epsilon;
    this->q0      = q0     ;
    this->qtau    = qtau   ;
    this->qmin    = qmin   ;
    this->lambda  = lambda ;
    this->nu      = nu     ;
 
    q = q0;
    P /= epsilon;
    gamma = 1.0;
 
    return;
}

References epsilon, gamma, lambda, nu, P, q, q0, qmin, and qtau.

Referenced by nnp::Training::setupTraining().

Here is the caller graph for this function:

◆ status()

string KalmanFilter::status ( std::size_t epoch ) const

virtual

Status report.

Parameters

[in] epoch Current epoch.

Returns: Line with current status information.

Implements nnp::Updater.

Definition at line 243 of file KalmanFilter.cpp.

{
 
    double Pasym = 0.5 * (P - P.transpose()).array().abs().mean();
    double Pdiag = P.diagonal().array().abs().sum();
    double Poffdiag = (P.array().abs().sum() - Pdiag)
                    / (sizeState * (sizeState - 1));
    Pdiag /= sizeState;
    double Kmean = K.array().abs().mean();
 
    string s = strpr("%10zu %10zu %16.8E %16.8E %16.8E %16.8E %16.8E",
                     epoch, numUpdates, Pdiag, Poffdiag, Pasym, Kmean, q);
    if (type == KT_STANDARD)
    {
        s += strpr(" %16.8E", eta);
    }
    else if (type == KT_FADINGMEMORY)
    {
        s += strpr(" %16.8E %16.8E", lambda, numUpdates * gamma);
    }
    s += '\n';
 
    return s;
}

References eta, gamma, K, KT_FADINGMEMORY, KT_STANDARD, lambda, numUpdates, P, q, nnp::Updater::sizeState, nnp::strpr(), and type.

Here is the call graph for this function:

◆ statusHeader()

vector< string > KalmanFilter::statusHeader ( ) const

virtual

Header for status report file.

Returns: Vector with header lines.

Implements nnp::Updater.

Definition at line 268 of file KalmanFilter.cpp.

{
    vector<string> header;
 
    vector<string> title;
    vector<string> colName;
    vector<string> colInfo;
    vector<size_t> colSize;
    title.push_back("Kalman filter status report.");
    colSize.push_back(10);
    colName.push_back("epoch");
    colInfo.push_back("Training epoch.");
    colSize.push_back(10);
    colName.push_back("nupdates");
    colInfo.push_back("Number of updates performed.");
    colSize.push_back(16);
    colName.push_back("Pdiag");
    colInfo.push_back("Mean of absolute diagonal values of error covariance "
                      "matrix P.");
    colSize.push_back(16);
    colName.push_back("Poffdiag");
    colInfo.push_back("Mean of absolute off-diagonal values of error "
                      "covariance matrix P.");
    colSize.push_back(16);
    colName.push_back("Pasym");
    colInfo.push_back("Asymmetry of P, i.e. mean of absolute values of "
                      "asymmetric part 0.5*(P - P^T).");
    colSize.push_back(16);
    colName.push_back("Kmean");
    colInfo.push_back("Mean of abolute compontents of Kalman gain matrix K.");
    colSize.push_back(16);
    colName.push_back("q");
    colInfo.push_back("Magnitude of process noise (= diagonal entries of Q).");
    if (type == KT_STANDARD)
    {
        colSize.push_back(16);
        colName.push_back("eta");
        colInfo.push_back("Learning rate.");
    }
    else if (type == KT_FADINGMEMORY)
    {
        colSize.push_back(16);
        colName.push_back("lambda");
        colInfo.push_back("Forgetting factor for fading memory KF.");
        colSize.push_back(16);
        colName.push_back("kgamma");
        colInfo.push_back("Forgetting gain k * gamma(k).");
    }
    header = createFileHeader(title, colSize, colName, colInfo);
 
    return header;
}

References nnp::createFileHeader(), KT_FADINGMEMORY, KT_STANDARD, and type.

Here is the call graph for this function:

◆ info()

vector< string > KalmanFilter::info ( ) const

virtual

Information about Kalman filter settings.

Returns: Vector with info lines.

Implements nnp::Updater.

Definition at line 321 of file KalmanFilter.cpp.

{
    vector<string> v;
 
    if (type == KT_STANDARD)
    {
        v.push_back(strpr("KalmanType::KT_STANDARD (%d)\n", type));
        v.push_back(strpr("sizeState       = %zu\n", sizeState));
        v.push_back(strpr("sizeObservation = %zu\n", sizeObservation));
        v.push_back(strpr("epsilon         = %12.4E\n", epsilon));
        v.push_back(strpr("q0              = %12.4E\n", q0     ));
        v.push_back(strpr("qtau            = %12.4E\n", qtau   ));
        v.push_back(strpr("qmin            = %12.4E\n", qmin   ));
        v.push_back(strpr("eta0            = %12.4E\n", eta0   ));
        v.push_back(strpr("etatau          = %12.4E\n", etatau ));
        v.push_back(strpr("etamax          = %12.4E\n", etamax ));
    }
    else if (type == KT_FADINGMEMORY)
    {
        v.push_back(strpr("KalmanType::KT_FADINGMEMORY (%d)\n", type));
        v.push_back(strpr("sizeState       = %zu\n", sizeState));
        v.push_back(strpr("sizeObservation = %zu\n", sizeObservation));
        v.push_back(strpr("epsilon         = %12.4E\n", epsilon));
        v.push_back(strpr("q0              = %12.4E\n", q0     ));
        v.push_back(strpr("qtau            = %12.4E\n", qtau   ));
        v.push_back(strpr("qmin            = %12.4E\n", qmin   ));
        v.push_back(strpr("lambda          = %12.4E\n", lambda));
        v.push_back(strpr("nu              = %12.4E\n", nu    ));
    }
    v.push_back(strpr("OpenMP threads used: %d\n", nbThreads()));
 
    return v;
}

References epsilon, eta0, etamax, etatau, KT_FADINGMEMORY, KT_STANDARD, lambda, nu, q0, qmin, qtau, sizeObservation, nnp::Updater::sizeState, nnp::strpr(), and type.

Here is the call graph for this function:

◆ getType()

KalmanFilter::KalmanType nnp::KalmanFilter::getType ( ) const

inline

Getter for type.

Definition at line 243 of file KalmanFilter.h.

{
    return type;
}

References type.

◆ getSizeObservation()

std::size_t nnp::KalmanFilter::getSizeObservation ( ) const

inline

Getter for sizeObservation.

Definition at line 248 of file KalmanFilter.h.

{
    return sizeObservation;
}

References sizeObservation.

◆ getNumUpdates()

std::size_t nnp::KalmanFilter::getNumUpdates ( ) const

inline

Getter for numUpdates.

Definition at line 253 of file KalmanFilter.h.

{
    return numUpdates;
}

References numUpdates.

◆ getEta()

double nnp::KalmanFilter::getEta ( ) const

inline

Getter for eta.

Definition at line 258 of file KalmanFilter.h.

{
    return eta;
}

References eta.

◆ getEpsilon()

double nnp::KalmanFilter::getEpsilon ( ) const

inline

Getter for epsilon.

Definition at line 263 of file KalmanFilter.h.

{
    return epsilon;
}

References epsilon.

◆ getQ0()

double nnp::KalmanFilter::getQ0 ( ) const

inline

Getter for q0.

Definition at line 268 of file KalmanFilter.h.

{
    return q0;
}

References q0.

◆ getQtau()

double nnp::KalmanFilter::getQtau ( ) const

inline

Getter for qtau.

Definition at line 273 of file KalmanFilter.h.

{
    return qtau;
}

References qtau.

◆ getQmin()

double nnp::KalmanFilter::getQmin ( ) const

inline

Getter for qmin.

Definition at line 278 of file KalmanFilter.h.

{
    return qmin;
}

References qmin.

◆ getLambda()

double nnp::KalmanFilter::getLambda ( ) const

inline

Getter for lambda.

Definition at line 283 of file KalmanFilter.h.

{
    return lambda;
}

References lambda.

◆ getNu()

double nnp::KalmanFilter::getNu ( ) const

inline

Getter for nu.

Definition at line 288 of file KalmanFilter.h.

{
    return nu;
}

References nu.

◆ getGamma()

double nnp::KalmanFilter::getGamma ( ) const

Getter for gamma.

Member Data Documentation

◆ type

KalmanType nnp::KalmanFilter::type

private

Kalman filter type.

Definition at line 196 of file KalmanFilter.h.

Referenced by getType(), info(), KalmanFilter(), status(), and statusHeader().

◆ sizeObservation

std::size_t nnp::KalmanFilter::sizeObservation

private

Size of observation (measurement) vector.

Definition at line 198 of file KalmanFilter.h.

Referenced by getSizeObservation(), info(), KalmanFilter(), setError(), setJacobian(), setSizeObservation(), and update().

◆ numUpdates

std::size_t nnp::KalmanFilter::numUpdates

private

Total number of updates performed.

Definition at line 200 of file KalmanFilter.h.

Referenced by getNumUpdates(), and status().

◆ epsilon

double nnp::KalmanFilter::epsilon

private

Error covariance initialization parameter \(\epsilon\).

Definition at line 202 of file KalmanFilter.h.

Referenced by getEpsilon(), info(), setParametersFadingMemory(), and setParametersStandard().

◆ q

double nnp::KalmanFilter::q

private

Process noise \(q\).

Definition at line 204 of file KalmanFilter.h.

Referenced by setParametersFadingMemory(), setParametersStandard(), and status().

◆ q0

double nnp::KalmanFilter::q0

private

Process noise initial value \(q_0\).

Definition at line 206 of file KalmanFilter.h.

Referenced by getQ0(), info(), setParametersFadingMemory(), and setParametersStandard().

◆ qtau

double nnp::KalmanFilter::qtau

private

Process noise exponential decay parameter \(q_{\tau}\).

Definition at line 208 of file KalmanFilter.h.

Referenced by getQtau(), info(), setParametersFadingMemory(), and setParametersStandard().

◆ qmin

double nnp::KalmanFilter::qmin

private

Process noise minimum value \(q_{\text{min}}\).

Definition at line 210 of file KalmanFilter.h.

Referenced by getQmin(), info(), setParametersFadingMemory(), and setParametersStandard().

◆ eta

double nnp::KalmanFilter::eta

private

Learning rate \(\eta\).

Definition at line 212 of file KalmanFilter.h.

Referenced by getEta(), setParametersStandard(), and status().

◆ eta0

double nnp::KalmanFilter::eta0

private

Learning rate initial value \(\eta_0\).

Definition at line 214 of file KalmanFilter.h.

Referenced by info(), and setParametersStandard().

◆ etatau

double nnp::KalmanFilter::etatau

private

Learning rate exponential increase parameter \(\eta_{\tau}\).

Definition at line 216 of file KalmanFilter.h.

Referenced by info(), and setParametersStandard().

◆ etamax

double nnp::KalmanFilter::etamax

private

Learning rate maximum value \(\eta_{\text{max}}\).

Definition at line 218 of file KalmanFilter.h.

Referenced by info(), and setParametersStandard().

◆ lambda

double nnp::KalmanFilter::lambda

private

Forgetting factor for fading memory Kalman filter.

Definition at line 220 of file KalmanFilter.h.

Referenced by getLambda(), info(), setParametersFadingMemory(), and status().

◆ nu

double nnp::KalmanFilter::nu

private

Parameter for fading memory Kalman filter.

Definition at line 222 of file KalmanFilter.h.

Referenced by getNu(), info(), and setParametersFadingMemory().

◆ gamma

double nnp::KalmanFilter::gamma

private

Forgetting gain factor gamma for fading memory Kalman filter.

Definition at line 224 of file KalmanFilter.h.

Referenced by setParametersFadingMemory(), and status().

◆ w

Eigen::Map<Eigen::VectorXd>* nnp::KalmanFilter::w

private

State vector.

Definition at line 226 of file KalmanFilter.h.

Referenced by KalmanFilter(), and setState().

◆ xi

Eigen::Map<Eigen::VectorXd const>* nnp::KalmanFilter::xi

private

Error vector.

Definition at line 228 of file KalmanFilter.h.

Referenced by KalmanFilter().

◆ H

Eigen::Map<Eigen::MatrixXd const>* nnp::KalmanFilter::H

private

Derivative matrix.

Definition at line 230 of file KalmanFilter.h.

Referenced by KalmanFilter().

◆ P

Eigen::MatrixXd nnp::KalmanFilter::P

private

Error covariance matrix.

Definition at line 232 of file KalmanFilter.h.

Referenced by KalmanFilter(), setParametersFadingMemory(), setParametersStandard(), and status().

◆ K

Eigen::MatrixXd nnp::KalmanFilter::K

private

Kalman gain matrix.

Definition at line 234 of file KalmanFilter.h.

Referenced by KalmanFilter(), and status().

◆ X

Eigen::MatrixXd nnp::KalmanFilter::X

private

Intermediate result X = P . H.

Definition at line 236 of file KalmanFilter.h.

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

/home/runner/work/n2p2/n2p2/src/libnnptrain/KalmanFilter.h
/home/runner/work/n2p2/n2p2/src/libnnptrain/KalmanFilter.cpp

Public Types

Public Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

Member Enumeration Documentation

◆ KalmanType

Constructor & Destructor Documentation

◆ KalmanFilter()

◆ ~KalmanFilter()

Member Function Documentation

◆ setSizeObservation()

◆ setState()

◆ setError() [1/2]

◆ setError() [2/2]

◆ setJacobian() [1/2]

◆ setJacobian() [2/2]

◆ update() [1/2]

◆ update() [2/2]

◆ setParametersStandard()

◆ setParametersFadingMemory()

◆ status()

◆ statusHeader()

◆ info()

◆ getType()

◆ getSizeObservation()

◆ getNumUpdates()

◆ getEta()

◆ getEpsilon()

◆ getQ0()

◆ getQtau()

◆ getQmin()

◆ getLambda()

◆ getNu()

◆ getGamma()

Member Data Documentation

◆ type

◆ sizeObservation

◆ numUpdates

◆ epsilon

◆ q

◆ q0

◆ qtau

◆ qmin

◆ eta

◆ eta0

◆ etatau

◆ etamax

◆ lambda

◆ nu

◆ gamma

◆ w

◆ xi

◆ H

◆ P

◆ K

◆ X