KF Class Reference

Kalman filter implementation, for a linear system. More...

#include <kf.h>

Public Member Functions
	KF ()
	Constructor, nothing happens here. More...
	~KF ()
	Destructor, nothing happens here. More...
void	InitSystem (const mat &A, const mat &B, const mat &H, const mat &Q, const mat &R)
	Define the system. More...
void	InitSystemState (const colvec &x0)
	Initialize the system states. Must be called after InitSystem. If not, called, system states are initialized to zero. More...
void	InitStateCovariance (const mat &P0)
	Initialize the state covariance. Must be called after InitSystem. If not called, covariance state is Initialized to an identity matrix. More...
void	Kalmanf (const colvec &u)
	Do Kalman filter iteration step-by-step while simulating the system. Simulating the system is done to calculate system states and outputs. More...
void	Kalmanf (const colvec &z, const colvec &u)
	Do Kalman filter 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. More...
colvec *	GetCurrentEstimatedOutput ()
	Get current estimated output. This is the filtered measurements, with less noise. More...

Detailed Description

Kalman filter implementation, for a linear system.

Constructor & Destructor Documentation

KF()

KF::KF

(

)

Constructor, nothing happens here.

~KF()

KF::~KF

(

)

Destructor, nothing happens here.

Member Function Documentation

GetCurrentEstimatedOutput()

colvec * KF::GetCurrentEstimatedOutput

(

)

Get current estimated output. This is the filtered measurements, with less noise.

Returns

Current estimated output $\hat{z}_k$

GetCurrentEstimatedState()

colvec * KF::GetCurrentEstimatedState

(

)

Get current estimated state.

Returns

Current estimated state $\hat{x}_k$

GetCurrentOutput()

colvec * KF::GetCurrentOutput

(

)

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

Returns

Current simulated output $z_k$

GetCurrentState()

colvec * KF::GetCurrentState

(

)

Get current simulated true state.

Returns

Current simulated state $x_k$

InitStateCovariance()

void KF::InitStateCovariance

(

const mat &

P0

)

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

Parameters

P0	Inital value for the state covariance

InitSystem()

void KF::InitSystem	(	const mat &	A,
		const mat &	B,
		const mat &	H,
		const mat &	Q,
		const mat &	R
	)

Define the system.

Parameters

A	System matrix
B	Input matrix
H	Output matrix
Q	Process noise covariance
R	Measurement noise covariance

InitSystemState()

void KF::InitSystemState

(

const colvec &

x0

)

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

Parameters

x0	Inital value for the system state

Kalmanf() [1/2]

void KF::Kalmanf

(

const colvec &

u

)

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

Parameters

u	tThe applied input to the system

Kalmanf() [2/2]

void KF::Kalmanf	(	const colvec &	z,
		const colvec &	u
	)

Do Kalman filter 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 values of the output from measurement
u	The applied input to the system

---

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

• kalman/kf.h
• kalman/kf.cpp