Kalman filter with constant-velocity system model. The system noise is automatically calculated from the maximum expected acceleration which can be set as a parameter. More...

#include <KalmanFilter.h>

Public Member Functions
virtual KalmanFilter *	clone () const
	Clone operator. More...

const pbl::Gaussian &	getGaussian () const
	Returns the Kalman state and covariance as Gaussian. More...

double	getLikelihood (const pbl::Gaussian &z) const
	Calculates the likelihood that the given measurement originates from the estimated Kalman state. More...

const pbl::Vector &	getState () const
	Returns the Kalman state (mean of the estimated Gaussian) More...

const pbl::Matrix &	getStateCovariance () const
	Returns the covariance of the Kalman state. More...

virtual void	init (const pbl::Gaussian &G)
	Initializes the Kalman filter according to the given Gaussian, i.e., the initial state and state covariance is set to the mean and covariance of the given Gaussian. More...

	KalmanFilter (int dim)
	Constructs a Kalman filter with specified state dimensionality. More...

	KalmanFilter (const KalmanFilter &orig)
	Copy constructor. More...

virtual void	propagate (const double &dt)
	Propagates the state of the Kalman filter according to the (constant- velocity) system model. More...

void	setMaxAcceleration (double a_max)
	Set the maximum expected acceleration. This parameter is used to determine the system noise in the propagate phase: a constant velocity is assumed, but the system noise takes into account that the the target may accelerate or deccelerate with the specified maximum acceleration. More...

void	update (const pbl::Gaussian &z)
	Updates the Kalman filter with the given (Gaussian) measurement. More...

virtual	~KalmanFilter ()
	Destructor. More...

Protected Attributes
double	a_max_

pbl::Gaussian	G_

pbl::Gaussian	G_small_

pbl::Matrix	H_

int	meas_dim_

int	state_dim_

Detailed Description

Kalman filter with constant-velocity system model. The system noise is automatically calculated from the maximum expected acceleration which can be set as a parameter.

Definition at line 46 of file KalmanFilter.h.

Constructor & Destructor Documentation

KalmanFilter::KalmanFilter ( int dim )

Constructs a Kalman filter with specified state dimensionality.

Parameters

dim	Dimensionality of the state

Definition at line 39 of file KalmanFilter.cpp.

KalmanFilter::KalmanFilter ( const KalmanFilter & orig )

Copy constructor.

Definition at line 43 of file KalmanFilter.cpp.

KalmanFilter::~KalmanFilter ( )

virtual

Destructor.

Definition at line 47 of file KalmanFilter.cpp.

Member Function Documentation

KalmanFilter * KalmanFilter::clone ( ) const

virtual

Clone operator.

Definition at line 50 of file KalmanFilter.cpp.

const pbl::Gaussian & KalmanFilter::getGaussian ( ) const

Returns the Kalman state and covariance as Gaussian.

Returns: The Kalman state and covariance as Gaussian

Definition at line 129 of file KalmanFilter.cpp.

double KalmanFilter::getLikelihood ( const pbl::Gaussian & z ) const

Calculates the likelihood that the given measurement originates from the estimated Kalman state.

Parameters

z	The Gaussian measurement

Returns

Definition at line 125 of file KalmanFilter.cpp.

const pbl::Vector & KalmanFilter::getState ( ) const

Returns the Kalman state (mean of the estimated Gaussian)

Returns: The Kalman state

Definition at line 133 of file KalmanFilter.cpp.

const pbl::Matrix & KalmanFilter::getStateCovariance ( ) const

Returns the covariance of the Kalman state.

Returns: The covariance of the Kalman state

Definition at line 137 of file KalmanFilter.cpp.

void KalmanFilter::init ( const pbl::Gaussian & G )

virtual

Initializes the Kalman filter according to the given Gaussian, i.e., the initial state and state covariance is set to the mean and covariance of the given Gaussian.

Parameters

G	The Gaussian to initialize with

Definition at line 54 of file KalmanFilter.cpp.

void KalmanFilter::propagate ( const double & dt )

virtual

Propagates the state of the Kalman filter according to the (constant- velocity) system model.

Parameters

dt	The time increment with which to propagate

Definition at line 64 of file KalmanFilter.cpp.

void KalmanFilter::setMaxAcceleration ( double a_max )

Set the maximum expected acceleration. This parameter is used to determine the system noise in the propagate phase: a constant velocity is assumed, but the system noise takes into account that the the target may accelerate or deccelerate with the specified maximum acceleration.

Parameters

a_max The maximum expected acceleration

Definition at line 141 of file KalmanFilter.cpp.

void KalmanFilter::update ( const pbl::Gaussian & z )

Updates the Kalman filter with the given (Gaussian) measurement.

Parameters

z	The Gaussian measurement

Definition at line 102 of file KalmanFilter.cpp.

Member Data Documentation

double KalmanFilter::a_max_

protected

Maximum expected acceleration

Definition at line 144 of file KalmanFilter.h.

pbl::Gaussian KalmanFilter::G_

protected

The estimated Kalman state as Gaussian (state mean and covariance)

Definition at line 135 of file KalmanFilter.h.

pbl::Gaussian KalmanFilter::G_small_

protected

Portion of the estimated Kalman state that contains only the measurement dimensions

Definition at line 138 of file KalmanFilter.h.

pbl::Matrix KalmanFilter::H_

protected

Observation model

Definition at line 141 of file KalmanFilter.h.

int KalmanFilter::meas_dim_

protected

Dimensionality of the measurements

Definition at line 129 of file KalmanFilter.h.

int KalmanFilter::state_dim_

protected

Dimensionality of the full state

Definition at line 132 of file KalmanFilter.h.

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

Public Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation