EKF2 Class Reference

Implemetation of the extended Kalman filter. This class needs to be derived. More...

#include <ekf2.h>

Public Member Functions
	EKF2 ()
	Constructor, nothing happens here. More...
	~EKF2 ()
	Destructur, nothing happens here. More...
void	InitSystem (int n_states, int n_outputs, const mat &Q, const mat &R)
	Tell me how many states and outputs you have! More...
virtual colvec	f (const colvec &x, const colvec &u)
	Define model of your system. More...
virtual colvec	h (const colvec &x)
	Define the output model of your system. More...
void	InitSystemState (const colvec &x0)
	Initialize the system states. Must be called after InitSystem. If not called, system state is initialized to zero. More...
void	InitSystemStateCovariance (const mat &P0)
	Initialize the system state covariance. Must be called after InitSystem. If not called, state covariance is initialized to an identity matrix. More...
void	EKalmanf (const colvec &u)
	Do the extended Kalman iteration step-by-step while simulating the system. Simulating the system is done to calculate system states and outputs. More...
void	EKalmanf (const colvec &z, const colvec &u)
	Do the extended Kalman iteration step-by-step without simulating the system. Use this if measurement is available and simulating the system is unnecessary. Here, true system states and system outputs do not matter. The only thing that matters is the estimated states. More...
colvec *	GetCurrentState ()
	Get current simulated true state. More...
colvec *	GetCurrentOutput ()
	Get current simulated true output. This is analogous to the measurements. More...
colvec *	GetCurrentEstimatedState ()
	Get current estimated state. This is analogous to the filtered measurements. More...
colvec *	GetCurrentEstimatedOutput ()
	Get current estimated output. More...

Protected Attributes
int	nStates_
	Number of the states. More...
int	nOutputs_
	Number of outputs. More...
colvec	x_
	State vector. More...
colvec	z_
	Output matrix. More...

Detailed Description

Implemetation of the extended Kalman filter. This class needs to be derived.

Constructor & Destructor Documentation

EKF2()

EKF2::EKF2

(

)

Constructor, nothing happens here.

~EKF2()

EKF2::~EKF2

(

)

Destructur, nothing happens here.

Member Function Documentation

EKalmanf() [1/2]

void EKF2::EKalmanf

(

const colvec &

u

)

Do the extended Kalman iteration step-by-step while simulating the system. Simulating the system is done to calculate system states and outputs.

Parameters

u	The applied input to the system

EKalmanf() [2/2]

void EKF2::EKalmanf	(	const colvec &	z,
		const colvec &	u
	)

Do the extended Kalman iteration step-by-step without simulating the system. Use this if measurement is available and simulating the system is unnecessary. Here, true system states and system outputs do not matter. The only thing that matters is the estimated states.

Parameters

z	The measurement outputs, this is a returned value
u	The applied input to the system

f()

colvec EKF2::f ( const colvec &  x, const colvec &  u  )

virtual

Define model of your system.

Parameters

x	System states
u	System inputs

GetCurrentEstimatedOutput()

colvec * EKF2::GetCurrentEstimatedOutput

(

)

Get current estimated output.

Returns

Current estimated output \(\hat{z}_k\)

GetCurrentEstimatedState()

colvec * EKF2::GetCurrentEstimatedState

(

)

Get current estimated state. This is analogous to the filtered measurements.

Returns

Current estimated state \(\hat{x}_k\)

GetCurrentOutput()

colvec * EKF2::GetCurrentOutput

(

)

Get current simulated true output. This is analogous to the measurements.

Returns

Current simulated output \(z_k\)

GetCurrentState()

colvec * EKF2::GetCurrentState

(

)

Get current simulated true state.

Returns

Current simulated state of the system \(x_k\)

h()

colvec EKF2::h ( const colvec &  x )

virtual

Define the output model of your system.

Parameters

x	System states

InitSystem()

void EKF2::InitSystem	(	int	n_states,
		int	n_outputs,
		const mat &	Q,
		const mat &	R
	)

Tell me how many states and outputs you have!

Parameters

n_states	Number of the elements on the input vector x
n_outputs	Number of the elements on the otput vector z
Q	Process noise covariance
R	Measurement noise covariance

InitSystemState()

void EKF2::InitSystemState

(

const colvec &

x0

)

Initialize the system states. Must be called after InitSystem. If not called, system state is initialized to zero.

Parameters

x0	Inital value for the system state

InitSystemStateCovariance()

void EKF2::InitSystemStateCovariance

(

const mat &

P0

)

Initialize the system state covariance. Must be called after InitSystem. If not called, state covariance is initialized to an identity matrix.

Parameters

P0	Inital value for the state covariance

Member Data Documentation

nOutputs_

int EKF2::nOutputs_

protected

Number of outputs.

nStates_

int EKF2::nStates_

protected

Number of the states.

x_

colvec EKF2::x_

protected

State vector.

z_

colvec EKF2::z_

protected

Output matrix.

---

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

• kalman/ekf2.h
• kalman/ekf2.cpp