jafar::rtslam::Perturbation Class Reference

Base class for all Gaussian perturbation vectors defined in the module rtslam. More...

---

Detailed Description

Base class for all Gaussian perturbation vectors defined in the module rtslam.

Author:

jsola

The Perturbation class is mainly a Gaussian. It represents discrete-time perturbation vectors.

In case the perturbation and perturbation values want to be specified in continuous-time, this class incorporates private members for storing the continuous values, and also methods for the conversion to discrete-time.

Definition at line 40 of file perturbation.hpp.

#include <perturbation.hpp>

Inheritance diagram for jafar::rtslam::Perturbation:

List of all members.

Public Member Functions
	Perturbation (const size_t _size)
virtual void	resize (size_t _size, bool preserve=true)
	Perturbation (const vec &p, const sym_mat &P)
	Perturbation (const vec &p, const vec &_std)
	Perturbation (const vec &p, const sym_mat &P, double dt)
	Continuous-time specified constructor.
	Perturbation (const vec &p, const vec &_std, double dt)
	Continuous-time specified constructor.
	Perturbation (const Gaussian &p)
void	clear ()
	Clear data, keep sizes and ranges.
template<class SymMat >
void	set_P_continuous (SymMat &_P_ct)
template<class V >
void	set_x_continuous (V &_x_ct)
void	set_std_continuous (const vec &_std)
	Set continuous-time covariance from standard deviation specification.
void	set_continuous (const vec &_x, const vec &_std)
	Set continuous-time perturbation.
void	set_continuous (const vec &_x, const sym_mat &_P)
	Set continuous-time perturbation.
void	set_deterministic (jblas::ind_array ia)
void	set_deterministic (ublas::range ia)
void	set_P_from_continuous (double _dt)
	Discrete perturbation from continuous specification.
void	set_P_from_continuous (sym_mat &_P_ct, double _dt)
	Discrete perturbation from continuous specification.
void	set_from_continuous (double _dt)
	Discrete perturbation from continuous specifications.
void	set_from_continuous (Gaussian &Pct, double _dt)
	Discrete perturbation from continuous specifications.
Private Attributes
vec	x_ct
	continuous-time perturbation vector
sym_mat	P_ct
	continuous-time covariances matrix
jblas::ind_array	ia_deterministic

---

Constructor & Destructor Documentation

jafar::rtslam::Perturbation::Perturbation	(	const vec &	p,
		const sym_mat &	P,
		double	dt
	)

Continuous-time specified constructor.

Parameters:

p	the perturbation mean in continuous time.
P	the perturbation covariance in continuous time
dt	the sampling time

jafar::rtslam::Perturbation::Perturbation	(	const vec &	p,
		const vec &	_std,
		double	dt
	)

Continuous-time specified constructor.

Parameters:

p	the perturbation mean in continuous time.
P	the perturbation std. deviation in continuous time
dt	the sampling time

---

Member Function Documentation

void jafar::rtslam::Perturbation::clear

(

void

)

[inline, virtual]

Clear data, keep sizes and ranges.

Clears the data of x_ and P_.

Reimplemented from jafar::rtslam::Gaussian.

Definition at line 95 of file perturbation.hpp.

References jafar::image::clear, and jafar::rtslam::Gaussian::clear().

void jafar::rtslam::Perturbation::set_from_continuous

(

double

_dt

)

[inline]

Discrete perturbation from continuous specifications.

• The deterministic values integrate with time normally, linearly with dt:
  • x = _x_ct * _dt
• The white, Gaussian random values integrate with the square root of dt. Their variance integrates linearly with dt:
  • x = _x_ct * sqrt(_dt)
  • P = _P_ct * _dt

This function takes mean and covariances from the internal variables of the class (which are often constant).

Parameters:

_dt	the time interval to integrate.

Definition at line 182 of file perturbation.hpp.

References JFR_ASSERT.

void jafar::rtslam::Perturbation::set_from_continuous	(	Gaussian &	Pct,
		double	_dt
	)		[inline]

Discrete perturbation from continuous specifications.

• The variance integrates linearly with dt:
  • P = Pct.P * _dt
• Therefore, the mean values integrate with time linearly with the square root of dt:
  • x = Pct.x * sqrt(_dt)

Parameters:

Pct	a continuous-time Gaussian process noise.
_dt	the time interval to integrate.

Definition at line 201 of file perturbation.hpp.

References JFR_ASSERT.

void jafar::rtslam::Perturbation::set_P_from_continuous

(

double

_dt

)

[inline]

Discrete perturbation from continuous specification.

• The white, Gaussian random values integrate with the square root of dt. Their variance integrates linearly with dt:
  • P = _P_ct * _dt

This function takes covariances from the internal variables of the class (which is often constant).

Parameters:

_dt	the time interval to integrate.

Definition at line 151 of file perturbation.hpp.

References JFR_ASSERT.

void jafar::rtslam::Perturbation::set_P_from_continuous	(	sym_mat &	_P_ct,
		double	_dt
	)		[inline]

Discrete perturbation from continuous specification.

• The white, Gaussian random values integrate with the square root of dt. Their variance integrates linearly with dt:
  • P = _P_ct * _dt

Parameters:

_P_ct	continuous-time perturbation covariances matrix.
_dt	the time interval to integrate.

Definition at line 165 of file perturbation.hpp.

References JFR_ASSERT.

void jafar::rtslam::Perturbation::set_std_continuous

(

const vec &

_std

)

[inline]

Set continuous-time covariance from standard deviation specification.

Parameters:

_std	a vector with the standard deviations.

Definition at line 116 of file perturbation.hpp.

References JFR_ASSERT.

---

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

• modules/rtslam/include/rtslam/perturbation.hpp