Package org.hipparchus.util

Interface UnscentedTransformProvider

All Known Implementing Classes:

AbstractUnscentedTransform, JulierUnscentedTransform, MerweUnscentedTransform

public interface UnscentedTransformProvider

Provider for unscented transform.

Since:

2.2

Method Summary

Modifier and Type	Method	Description
default RealMatrix	getUnscentedCovariance(RealVector[] sigmaPoints, RealVector meanState)	Computes the unscented covariance matrix from a weighted mean state and a set of sigma points.
default RealVector	getUnscentedMeanState(RealVector[] sigmaPoints)	Computes a weighted mean state from a given set of sigma points.
RealVector	getWc()	Get the covariance weights.
RealVector	getWm()	Get the mean weights.
default Pair<RealVector,RealMatrix>	inverseUnscentedTransform(RealVector[] sigmaPoints)	Perform the inverse unscented transform from an array of sigma points.
RealVector[]	unscentedTransform(RealVector state, RealMatrix covariance)	Perform the unscented transform from a state and its covariance.

Method Detail

unscentedTransform

RealVector[] unscentedTransform(RealVector state,
                                RealMatrix covariance)

Perform the unscented transform from a state and its covariance.

Parameters:

state - process state

covariance - covariance associated with the process state

Returns:

an array containing the sigma points of the unscented transform

getUnscentedMeanState

default RealVector getUnscentedMeanState(RealVector[] sigmaPoints)

Computes a weighted mean state from a given set of sigma points.

This method can be used for computing both the mean state and the mean measurement in an Unscented Kalman filter.

It corresponds to Equation 17 of "Wan, E. A., & Van Der Merwe, R. The unscented Kalman filter for nonlinear estimation"

Parameters:

sigmaPoints - input samples

Returns:

weighted mean state

getUnscentedCovariance

default RealMatrix getUnscentedCovariance(RealVector[] sigmaPoints,
                                          RealVector meanState)

Computes the unscented covariance matrix from a weighted mean state and a set of sigma points.

This method can be used for computing both the predicted state covariance matrix and the innovation covariance matrix in an Unscented Kalman filter.

It corresponds to Equation 18 of "Wan, E. A., & Van Der Merwe, R. The unscented Kalman filter for nonlinear estimation"

Parameters:

sigmaPoints - input sigma points

meanState - weighted mean state

Returns:

the unscented covariance matrix

inverseUnscentedTransform

default Pair<RealVector,RealMatrix> inverseUnscentedTransform(RealVector[] sigmaPoints)

Perform the inverse unscented transform from an array of sigma points.

Parameters:

sigmaPoints - array containing the sigma points of the unscented transform

Returns:

mean state and associated covariance

getWc

RealVector getWc()

Get the covariance weights.

Returns:

the covariance weights

getWm

RealVector getWm()

Get the mean weights.

Returns:

the mean weights