Package org.apache.commons.math3.filter

Class DefaultProcessModel

java.lang.Object

org.apache.commons.math3.filter.DefaultProcessModel

All Implemented Interfaces:

ProcessModel

public class DefaultProcessModel
extends java.lang.Object
implements ProcessModel

Default implementation of a ProcessModel for the use with a KalmanFilter.

Since:

3.0

Field Summary

Fields

Modifier and Type	Field	Description
private RealMatrix	controlMatrix	The control matrix, used to integrate a control input into the state estimation.
private RealMatrix	initialErrorCovMatrix	The initial error covariance matrix of the observed process.
private RealVector	initialStateEstimateVector	The initial state estimation of the observed process.
private RealMatrix	processNoiseCovMatrix	The process noise covariance matrix.
private RealMatrix	stateTransitionMatrix	The state transition matrix, used to advance the internal state estimation each time-step.

Constructor Summary

Constructors

Constructor	Description
DefaultProcessModel(double[][] stateTransition, double[][] control, double[][] processNoise)	Create a new ProcessModel, taking double arrays as input parameters.
DefaultProcessModel(double[][] stateTransition, double[][] control, double[][] processNoise, double[] initialStateEstimate, double[][] initialErrorCovariance)	Create a new ProcessModel, taking double arrays as input parameters.
DefaultProcessModel(RealMatrix stateTransition, RealMatrix control, RealMatrix processNoise, RealVector initialStateEstimate, RealMatrix initialErrorCovariance)	Create a new ProcessModel, taking double arrays as input parameters.

Method Summary

Modifier and Type	Method	Description
RealMatrix	getControlMatrix()	Returns the control matrix.
RealMatrix	getInitialErrorCovariance()	Returns the initial error covariance matrix.
RealVector	getInitialStateEstimate()	Returns the initial state estimation vector.
RealMatrix	getProcessNoise()	Returns the process noise matrix.
RealMatrix	getStateTransitionMatrix()	Returns the state transition matrix.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

stateTransitionMatrix

private RealMatrix stateTransitionMatrix

The state transition matrix, used to advance the internal state estimation each time-step.

controlMatrix

private RealMatrix controlMatrix

The control matrix, used to integrate a control input into the state estimation.

processNoiseCovMatrix

private RealMatrix processNoiseCovMatrix

The process noise covariance matrix.

initialStateEstimateVector

private RealVector initialStateEstimateVector

The initial state estimation of the observed process.

initialErrorCovMatrix

private RealMatrix initialErrorCovMatrix

The initial error covariance matrix of the observed process.

Constructor Detail

DefaultProcessModel

public DefaultProcessModel(double[][] stateTransition,
                           double[][] control,
                           double[][] processNoise,
                           double[] initialStateEstimate,
                           double[][] initialErrorCovariance)
                    throws NullArgumentException,
                           NoDataException,
                           DimensionMismatchException

Create a new ProcessModel, taking double arrays as input parameters.

Parameters:

stateTransition - the state transition matrix

control - the control matrix

processNoise - the process noise matrix

initialStateEstimate - the initial state estimate vector

initialErrorCovariance - the initial error covariance matrix

Throws:

NullArgumentException - if any of the input arrays is null

NoDataException - if any row / column dimension of the input matrices is zero

DimensionMismatchException - if any of the input matrices is non-rectangular

DefaultProcessModel

public DefaultProcessModel(double[][] stateTransition,
                           double[][] control,
                           double[][] processNoise)
                    throws NullArgumentException,
                           NoDataException,
                           DimensionMismatchException

Create a new ProcessModel, taking double arrays as input parameters.

The initial state estimate and error covariance are omitted and will be initialized by the KalmanFilter to default values.

Parameters:

stateTransition - the state transition matrix

control - the control matrix

processNoise - the process noise matrix

Throws:

NullArgumentException - if any of the input arrays is null

NoDataException - if any row / column dimension of the input matrices is zero

DimensionMismatchException - if any of the input matrices is non-rectangular

DefaultProcessModel

public DefaultProcessModel(RealMatrix stateTransition,
                           RealMatrix control,
                           RealMatrix processNoise,
                           RealVector initialStateEstimate,
                           RealMatrix initialErrorCovariance)

Create a new ProcessModel, taking double arrays as input parameters.

Parameters:

stateTransition - the state transition matrix

control - the control matrix

processNoise - the process noise matrix

initialStateEstimate - the initial state estimate vector

initialErrorCovariance - the initial error covariance matrix

Method Detail

getStateTransitionMatrix

public RealMatrix getStateTransitionMatrix()

Returns the state transition matrix.

Specified by:

getStateTransitionMatrix in interface ProcessModel

Returns:

the state transition matrix

getControlMatrix

public RealMatrix getControlMatrix()

Returns the control matrix.

Specified by:

getControlMatrix in interface ProcessModel

Returns:

the control matrix

getProcessNoise

public RealMatrix getProcessNoise()

Returns the process noise matrix. This method is called by the KalmanFilter every prediction step, so implementations of this interface may return a modified process noise depending on the current iteration step.

Specified by:

getProcessNoise in interface ProcessModel

Returns:

the process noise matrix

See Also:

KalmanFilter.predict(), KalmanFilter.predict(double[]), KalmanFilter.predict(RealVector)

getInitialStateEstimate

public RealVector getInitialStateEstimate()

Returns the initial state estimation vector.

Note: if the return value is zero, the Kalman filter will initialize the state estimation with a zero vector.

Specified by:

getInitialStateEstimate in interface ProcessModel

Returns:

the initial state estimation vector

getInitialErrorCovariance

public RealMatrix getInitialErrorCovariance()

Returns the initial error covariance matrix.

Note: if the return value is zero, the Kalman filter will initialize the error covariance with the process noise matrix.

Specified by:

getInitialErrorCovariance in interface ProcessModel

Returns:

the initial error covariance matrix