navlib::ExtendedKalmanFilter< T > Class Template Reference

A Kalman Filter that is able to operate on non-linear systems by linearizing the dynamics about the current state estimate. More...

#include <navmath.h>

List of all members.

Public Types
typedef boost::function < ublas::vector< T > ublas::vector< T >) >	PredictFun
	Prediction function.
typedef boost::function < ublas::vector< T > ublas::vector< T >) >	ObserveFun
	Observation function.
Public Member Functions
	ExtendedKalmanFilter (ublas::vector< T > x0, T kP=1.0e-9)
	ExtendedKalmanFilter (ublas::vector< T > x0, ublas::matrix< T > P0)
void	predict (PredictFun f, ublas::matrix< T > Q, T tol=1.0e-9)
	Predicts the next state.
void	correct (ObserveFun h, ublas::vector< T > z, ublas::matrix< T > R, T tol=1.0e-9)
	Corrects for errors between the observation and current state estimate.
ublas::vector< T >	getState () const
	Returns the current state estimate.

---

Detailed Description

template<typename T = double>
class navlib::ExtendedKalmanFilter< T >

A Kalman Filter that is able to operate on non-linear systems by linearizing the dynamics about the current state estimate.

---

Constructor & Destructor Documentation

template<typename T = double>

navlib::ExtendedKalmanFilter< T >::ExtendedKalmanFilter	(	ublas::vector< T >	x0,
		T	kP = 1.0e-9
	)			[inline]

Parameters:

	x0	initial state estimate.
	kP	small value used to generate the state uncertainty matrix.

template<typename T = double>

navlib::ExtendedKalmanFilter< T >::ExtendedKalmanFilter	(	ublas::vector< T >	x0,
		ublas::matrix< T >	P0
	)			[inline]

Parameters:

	x0	initial state estimate
	P0	initial state uncertainty matrix

---

Member Function Documentation

template<typename T = double>

void navlib::ExtendedKalmanFilter< T >::predict	(	PredictFun	f,
		ublas::matrix< T >	Q,
		T	tol = 1.0e-9
	)			[inline]

Predicts the next state.

Parameters:

	f	prediction function
	Q	state noise covariance matrix
	tol	small value used to calculate the Jacobian

template<typename T = double>

void navlib::ExtendedKalmanFilter< T >::correct	(	ObserveFun	h,
		ublas::vector< T >	z,
		ublas::matrix< T >	R,
		T	tol = 1.0e-9
	)			[inline]

Corrects for errors between the observation and current state estimate.

Parameters:

	h	observation function
	z	measurement
	R	measuement noise covariance matrix

Referenced by navlib::EkfIdentifier::correct().

---

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

• navmath.h