jafar::filter::BlockObserveModel Class Reference

Non linear observation model defined by blocks (two blocks). More...

---

Detailed Description

Non linear observation model defined by blocks (two blocks).

To define your own model you must derive this class. This model is well suited for feature observation model in the SLAM problem.

Definition at line 213 of file observeModel.hpp.

#include <observeModel.hpp>

Inheritance diagram for jafar::filter::BlockObserveModel:

List of all members.

Public Member Functions
	BlockObserveModel (std::size_t sizeObs_, std::size_t sizeState1_, std::size_t sizeState2_)
	BlockObserveModel (std::size_t sizeObs_, std::size_t sizeState1_, std::size_t sizeState2_, std::size_t sizeInnovation_, std::size_t sizePrediction_)
std::size_t	sizeState1 () const
std::size_t	sizeState2 () const
virtual jblas::vec const &	predictObservation (const jblas::vec_range &x1_, const jblas::vec_range &x2_)=0
	Predict observation for state x_.
virtual jblas::vec const &	computeInnovation (jblas::vec const &y_, jblas::vec const &zPred_)
	By default, simply do the substraction.
virtual void	predictObservationJac (const jblas::vec_range &x1_, const jblas::vec_range &x2_)=0
Public Attributes
jblas::mat	Jobs1
	Jacobian of the observe function with respect to block1.
jblas::mat	Jobs2
	Jacobian of the observe function with respect to block2.
Protected Attributes
std::size_t	_sizeState1
std::size_t	_sizeState2

---

Member Function Documentation

virtual jblas::vec const& jafar::filter::BlockObserveModel::computeInnovation	(	jblas::vec const &	y_,
		jblas::vec const &	zPred_
	)		[inline, virtual]

By default, simply do the substraction.

Reimplementation of that function are allowed to adjust the values of Jobs1 and Jobs2.

Parameters:

y_	observation
zPred_	prediction

Reimplemented in jafar::slam::StereoImagePluckerFeatureObserveModel, jafar::slam::ScaledHomogeneousImagePluckerFeatureObserveModel, jafar::slam::RhoThetaImagePluckerFeatureObserveModel, jafar::slam::ImageEuclideanPluckerFeatureObserveModel, jafar::slam::ImageAnchorPluckerFeatureObserveModel, jafar::slam::ImageSegInvDepthFeatureObserveModel, jafar::slam::BearingPointFeatureObserveModel, jafar::slam::PolarPointFeatureObserveModel, jafar::slam::ImageSegInvDepthInovFeatureObserveModel, jafar::slam::BearingPointInfFeatureObserveModel, and jafar::slam::BearingPointInvDepthFeatureObserveModel.

Definition at line 270 of file observeModel.hpp.

References JFR_PRECOND, and jafar::filter::BaseObserveModel::z.

virtual jblas::vec const& jafar::filter::BlockObserveModel::predictObservation	(	const jblas::vec_range &	x1_,
		const jblas::vec_range &	x2_
	)		[pure virtual]

Predict observation for state x_.

In order to minimize memory allocation, z vector is allocated once, predictObservation() implementation should look like:

   const vec& predictObservation(const range_vec& x1_, 
                                 const range_vec& x2_) {
     JFR_PRECOND(x1_.size() == _sizeState1, message);
     JFR_PRECOND(x2_.size() == _sizeState2, message);
     // compute z function of x1_ and x2_
     return z;
   }

Implemented in jafar::slam::BaseFeatureObserveModel.

---

Member Data Documentation

jblas::mat jafar::filter::BlockObserveModel::Jobs1

Jacobian of the observe function with respect to block1.

Warning:

do not forget to call computeJac() before using Jobs1

Definition at line 226 of file observeModel.hpp.

jblas::mat jafar::filter::BlockObserveModel::Jobs2

Jacobian of the observe function with respect to block2.

Warning:

do not forget to call computeJac() before using Jobs2

Definition at line 232 of file observeModel.hpp.

---

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

• modules/filter/include/filter/observeModel.hpp