LinearAnalyticMeasurementModelGaussianUncertainty_Implicit Class Reference

Class for linear analytic measurementmodels with additive gaussian noise. More...

#include <linearanalyticmeasurementmodel_gaussianuncertainty_implicit.h>

Inheritance diagram for LinearAnalyticMeasurementModelGaussianUncertainty_Implicit:

LinearAnalyticMeasurementModelGaussianUncertainty AnalyticMeasurementModelGaussianUncertainty MeasurementModel< MatrixWrapper::ColumnVector, MatrixWrapper::ColumnVector >

List of all members.

Public Member Functions

 LinearAnalyticMeasurementModelGaussianUncertainty_Implicit (LinearAnalyticConditionalGaussian *pdf)
 Constructor.
 LinearAnalyticMeasurementModelGaussianUncertainty_Implicit ()
 Constructor.
virtual ~LinearAnalyticMeasurementModelGaussianUncertainty_Implicit ()
 Destructor.
virtual const
MatrixWrapper::ColumnVector
fGet () const =0
virtual const int TypeGet () const =0
virtual MatrixWrapper::MatrixdfGet (int number)=0
virtual MatrixWrapper::Matrix df_dxGet (const MatrixWrapper::ColumnVector &z, const MatrixWrapper::ColumnVector &x)=0
 Returns H-matrix calculated with measurement z and state x.
virtual MatrixWrapper::Matrixdf_dzGet (const MatrixWrapper::ColumnVector &z, const MatrixWrapper::ColumnVector &x)=0
 Returns D-matrix calculated with measurement z and state x.
virtual MatrixWrapper::ColumnVector PredictionGet (const MatrixWrapper::ColumnVector &z, const MatrixWrapper::ColumnVector &x)=0
 Return a prediction for the mean of the noise on the linear measurement equation, calculated with measurements z and state x.
virtual MatrixWrapper::ColumnVector ExpectedValueGet ()=0
 Return a prediction for the mean of the noise on the linear measurement equation, using the current x and z.
virtual
MatrixWrapper::SymmetricMatrix & 
CovarianceGet ()=0
 Returns covariance of the noise on the linearised measurement model evaluated using measurements z and states x.
virtual
MatrixWrapper::SymmetricMatrix 
CovarianceGet (const MatrixWrapper::ColumnVector &z, const MatrixWrapper::ColumnVector &x)=0
 Returns covariance of the noise on the linearised measurement model evaluated using current z and states x.
virtual void Calculate (const MatrixWrapper::ColumnVector &x, const MatrixWrapper::ColumnVector &z, const MatrixWrapper::Matrix &R)=0
virtual const
MatrixWrapper::Matrix
SRCovariance () const =0
 Returns square root of the covariance of the measurements z.
virtual const int & Is_Identity () const =0
 Returns 1 if D-matrix equals the identity matrix else 0.
void HSet (const MatrixWrapper::Matrix &h)
 Set Matrix H.
void JSet (const MatrixWrapper::Matrix &j)
 Set Matrix J.
const MatrixWrapper::MatrixHGet () const
 Get Matrix H.
const MatrixWrapper::MatrixJGet () const
 Get Matrix J.
int MeasurementSizeGet () const
 Get Measurement Size.
bool SystemWithoutSensorParams () const
 Number of Conditional Arguments.
ConditionalPdf
< MatrixWrapper::ColumnVector,
MatrixWrapper::ColumnVector > * 
MeasurementPdfGet ()
 Get the MeasurementPDF.
void MeasurementPdfSet (ConditionalPdf< MatrixWrapper::ColumnVector, MatrixWrapper::ColumnVector > *pdf)
 Set the MeasurementPDF.
MatrixWrapper::ColumnVector Simulate (const MatrixWrapper::ColumnVector &x, const MatrixWrapper::ColumnVector &s, int sampling_method=DEFAULT, void *sampling_args=NULL)
 Simulate the Measurement, given a certain state, and an input.
MatrixWrapper::ColumnVector Simulate (const MatrixWrapper::ColumnVector &x, int sampling_method=DEFAULT, void *sampling_args=NULL)
 Simulate the system (no input system).
Probability ProbabilityGet (const MatrixWrapper::ColumnVector &z, const MatrixWrapper::ColumnVector &x, const MatrixWrapper::ColumnVector &s)
 Get the probability of a certain measurement.
Probability ProbabilityGet (const MatrixWrapper::ColumnVector &z, const MatrixWrapper::ColumnVector &x)
 Get the probability of a certain measurement.

Protected Attributes

ConditionalPdf
< MatrixWrapper::ColumnVector,
MatrixWrapper::ColumnVector > * 
_MeasurementPdf
 ConditionalPdf representing $ P(Z_k | X_{k}, U_{k}) $.
bool _systemWithoutSensorParams
 System with no sensor params??


Detailed Description

Class for linear analytic measurementmodels with additive gaussian noise.

This class represents all measurement models of the form

\[ 0 = f (x,z) \]

as a linear measurement model with virtual measurement z_k^{virtual}

\[ z_k^{virtual} = H(x_k,z_k) \times x_k + N(\mu(x_{k},z_k) ,\Sigma(x_k,z_k)) \]

Definition at line 37 of file linearanalyticmeasurementmodel_gaussianuncertainty_implicit.h.


Constructor & Destructor Documentation

Constructor.

Parameters:
pdf Conditional pdf, with Gaussian uncertainty


Member Function Documentation

virtual MatrixWrapper::SymmetricMatrix CovarianceGet ( const MatrixWrapper::ColumnVector z,
const MatrixWrapper::ColumnVector x 
) [pure virtual]

Returns covariance of the noise on the linearised measurement model evaluated using current z and states x.

The linearised measurement equation look like:

\[ z_k^{virtual} = H(x_{k},z_k) \times x_k + N(\mu(x_{k},z_k) ,\Sigma(x_k,z_k)) \]

with noise

\[ =N(\mu(x_{k},z_k), \Sigma(x_k,z_k))\]

and covariance

\[ \Sigma(x_k,z_k)= D(x_k,z_k)*R*D(x_k,z_k)' \]

and R the noise on the measurements z .

Reimplemented from LinearAnalyticMeasurementModelGaussianUncertainty.

virtual MatrixWrapper::SymmetricMatrix& CovarianceGet (  )  [pure virtual]

Returns covariance of the noise on the linearised measurement model evaluated using measurements z and states x.

The linearised measurement equation look like:

\[ z_k^{virtual} = H(x_{k},z_k) \times x_k + N(\mu(x_{k},z_k) ,\Sigma(x_k,z_k)) \]

with noise

\[ =N(\mu(x_{k},z_k), \Sigma(x_k,z_k))\]

and covariance

\[ \Sigma(x_k,z_k)= D(x_k,z_k)*R*D(x_k,z_k)' \]

and R the noise on the measurements z .

virtual MatrixWrapper::Matrix df_dxGet ( const MatrixWrapper::ColumnVector z,
const MatrixWrapper::ColumnVector x 
) [pure virtual]

Returns H-matrix calculated with measurement z and state x.

\[ H = \frac{df}{dx} \mid_{ z, x} \]

used to determine the covariance of noise on the linear measurement equation

Parameters:
z The value of the input in which the derivate is evaluated
x The value in the state in which the derivate is evaluated

Reimplemented from LinearAnalyticMeasurementModelGaussianUncertainty.

virtual MatrixWrapper::Matrix& df_dzGet ( const MatrixWrapper::ColumnVector z,
const MatrixWrapper::ColumnVector x 
) [pure virtual]

Returns D-matrix calculated with measurement z and state x.

\[ D = \frac{df}{dz} \mid_{ z, x} \]

used to determine the covariance of noise on the linear measurement equation

Parameters:
z The value of the input in which the derivate is evaluated
x The value in the state in which the derivate is evaluated

void HSet ( const MatrixWrapper::Matrix h  )  [inherited]

Set Matrix H.

This can be particularly useful for time-varying systems

Parameters:
h Matrix H

void JSet ( const MatrixWrapper::Matrix j  )  [inherited]

Set Matrix J.

This can be particularly useful for time-varying systems

Parameters:
j Matrix J

void MeasurementPdfSet ( ConditionalPdf< MatrixWrapper::ColumnVector , MatrixWrapper::ColumnVector > *  pdf  )  [inherited]

Set the MeasurementPDF.

Parameters:
pdf a pointer to the measurement pdf

Probability ProbabilityGet ( const MatrixWrapper::ColumnVector z,
const MatrixWrapper::ColumnVector x 
) [inherited]

Get the probability of a certain measurement.

(measurement independent of input) gived a certain state and input

Parameters:
z the measurement value
x x current state of the system
Returns:
the "probability" of the measurement

Probability ProbabilityGet ( const MatrixWrapper::ColumnVector z,
const MatrixWrapper::ColumnVector x,
const MatrixWrapper::ColumnVector s 
) [inherited]

Get the probability of a certain measurement.

given a certain state and input

Parameters:
z the measurement value
x current state of the system
s the sensor param value
Returns:
the "probability" of the measurement

MatrixWrapper::ColumnVector Simulate ( const MatrixWrapper::ColumnVector x,
int  sampling_method = DEFAULT,
void *  sampling_args = NULL 
) [inherited]

Simulate the system (no input system).

Parameters:
x current state of the system
Returns:
State where we arrive by simulating the measurement model
Note:
Maybe the return value would better be a Sample<StateVar> instead of a StateVar
Parameters:
sampling_method the sampling method to be used while sampling from the Conditional Pdf describing the system (if not specified = DEFAULT)
sampling_args Sometimes a sampling method can have some extra parameters (eg mcmc sampling)

MatrixWrapper::ColumnVector Simulate ( const MatrixWrapper::ColumnVector x,
const MatrixWrapper::ColumnVector s,
int  sampling_method = DEFAULT,
void *  sampling_args = NULL 
) [inherited]

Simulate the Measurement, given a certain state, and an input.

Parameters:
x current state of the system
s sensor parameter
Returns:
Measurement generated by simulating the measurement model
Parameters:
sampling_method the sampling method to be used while sampling from the Conditional Pdf describing the system (if not specified = DEFAULT)
sampling_args Sometimes a sampling method can have some extra parameters (eg mcmc sampling)
Note:
Maybe the return value would better be a Sample<StateVar> instead of a StateVar


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

Generated on Thu Mar 24 16:57:12 2011 for Bayesian Filtering Library by  doxygen 1.5.9