Detailed Description

template<class Pose>
class gtsam::WnoaMotionFactor< Pose >

WNOA (White Noise on Acceleration) motion prior factor.

This factor implements the WNOA motion prior between two states (pose and velocity at times t_k and t_{k+1}). It provides residuals and Jacobians consistent with the WNOA continuous-time prior when discretized over the timestep deltaT.

Template parameter Pose is expected to be a GTSAM pose type (e.g. Pose2, Pose3) or a vector-space pose; the factor supports both Lie group and vector-space pose representations.

The factor's ordering of keys is: pose_k, vel_k, pose_kp1, vel_kp1.

Template Parameters

Pose	Pose group/type (Pose2, Pose3, etc.)

Inheritance diagram for gtsam::WnoaMotionFactor< Pose >:

Public Member Functions
	WnoaMotionFactor (const StateData &state_k, const StateData &state_kp1, const VectorN &q_psd_diag)
	Construct a WNOA motion factor from two StateData entries.
	WnoaMotionFactor (Key poseKey0, Key velKey0, Key poseKey1, Key velKey1, const double deltaT, const VectorN &q_psd_diag)
	Construct a WNOA factor given explicit keys and timestep.
Vector	evaluateError (const Pose &p1, const Velocity &v1, const Pose &p2, const Velocity &v2, OptionalMatrixType Hp1, OptionalMatrixType Hv1, OptionalMatrixType Hp2, OptionalMatrixType Hv2) const override
	Evaluate the WNOA factor residual and optional Jacobians.
virtual Vector	evaluateError (const ValueTypes &... x, OptionalMatrixTypeT< ValueTypes >... H) const =0
	Override evaluateError to finish implementing an n-way factor.
Vector	evaluateError (const ValueTypes &... x, MatrixTypeT< ValueTypes > &... H) const
	If all the optional arguments are matrices then redirect the call to the one which takes pointers.
Vector	evaluateError (const ValueTypes &... x) const
	No-Jacobians requested function overload.
AreAllMatrixRefs< Vector, OptionalJacArgs... >	evaluateError (const ValueTypes &... x, OptionalJacArgs &&... H) const
	Some (but not all) optional Jacobians are omitted (function overload) and the jacobians are l-value references to matrices.
Testable
void	print (const std::string &s="", const KeyFormatter &keyFormatter=DefaultKeyFormatter) const override
	print
bool	equals (const NonlinearFactor &expected, double tol=1e-9) const override
	equals
Public Member Functions inherited from gtsam::NoiseModelFactorT< Vector, ValueTypes... >
Key	key () const
	Returns a key.
Vector	unwhitenedError (const Values &x, std::vector< Matrix > &H) const
	support taking in the actual vector instead of the pointer as well to get access to this version of the function from derived classes one will need to use the "using" keyword and specify that like this: public: using NoiseModelFactor::unwhitenedError;
	NoiseModelFactorT ()
	Default Constructor for I/O.
	NoiseModelFactorT (const SharedNoiseModel &noiseModel, KeyType< ValueTypes >... keys)
	Constructor.
	NoiseModelFactorT (const SharedNoiseModel &noiseModel, CONTAINER keys)
	Constructor.
virtual Vector	evaluateError (const ValueTypes &... x, OptionalMatrixTypeT< ValueTypes >... H) const=0
	Override evaluateError to finish implementing an n-way factor.
Vector	evaluateError (const ValueTypes &... x, MatrixTypeT< ValueTypes > &... H) const
	If all the optional arguments are matrices then redirect the call to the one which takes pointers.
Vector	evaluateError (const ValueTypes &... x) const
	No-Jacobians requested function overload.
AreAllMatrixRefs< Vector, OptionalJacArgs... >	evaluateError (const ValueTypes &... x, OptionalJacArgs &&... H) const
	Some (but not all) optional Jacobians are omitted (function overload) and the jacobians are l-value references to matrices.
AreAllMatrixPtrs< Vector, OptionalJacArgs... >	evaluateError (const ValueTypes &... x, OptionalJacArgs &&... H) const
	Some (but not all) optional Jacobians are omitted (function overload) and the jacobians are pointers to matrices.
Key	key1 () const
Key	key2 () const
Key	key3 () const
Key	key4 () const
Key	key5 () const
Key	key6 () const
std::shared_ptr< GaussianFactor >	linearize (const Values &values) const override
	Linearize factors whose error and argument dimensions are all fixed to an arbitrary-arity FixedJacobianFactor.
Vector	unwhitenedError (const Values &x, OptionalMatrixVecType H=nullptr) const override
	This implements the unwhitenedError virtual function by calling the n-key specific version of evaluateError, which is pure virtual so must be implemented in the derived class.
Public Member Functions inherited from gtsam::NoiseModelFactor
	NoiseModelFactor ()
	Default constructor for I/O only.
	~NoiseModelFactor () override
	Destructor.
template<typename CONTAINER>
	NoiseModelFactor (const SharedNoiseModel &noiseModel, const CONTAINER &keys)
	Constructor.
size_t	dim () const override
	get the dimension of the factor (number of rows on linearization)
const SharedNoiseModel &	noiseModel () const
	access to the noise model
Vector	unwhitenedError (const Values &x, std::vector< Matrix > &H) const
	support taking in the actual vector instead of the pointer as well to get access to this version of the function from derived classes one will need to use the "using" keyword and specify that like this: public: using NoiseModelFactor::unwhitenedError;
Vector	whitenedError (const Values &c) const
	Vector of errors, whitened This is the raw error, i.e., i.e.
Vector	unweightedWhitenedError (const Values &c) const
	Vector of errors, whitened, but unweighted by any loss function.
double	weight (const Values &c) const
	Compute the effective weight of the factor from the noise model.
double	error (const Values &c) const override
	Calculate the error of the factor.
shared_ptr	cloneWithNewNoiseModel (const SharedNoiseModel newNoise) const
	Creates a shared_ptr clone of the factor with a new noise model.
double	error (const HybridValues &c) const override
	All factor types need to implement an error function.
	NonlinearFactor ()
	Default constructor for I/O only.
template<typename CONTAINER>
	NonlinearFactor (const CONTAINER &keys)
	Constructor from a collection of the keys involved in this factor.
double	error (const HybridValues &c) const override
	All factor types need to implement an error function.
virtual bool	active (const Values &c) const
	Checks whether a factor should be used based on a set of values.
virtual void	qcqpFactors (NonlinearFactorGraph costs, NonlinearEqualityConstraints constraints, size_t columnDimension=1) const
	Add this factor's QCQP cost and constraints over matrix-valued QCQP variables with the given column dimension.
virtual shared_ptr	clone () const
	Creates a shared_ptr clone of the factor - needs to be specialized to allow for subclasses.
virtual shared_ptr	rekey (const std::map< Key, Key > &rekey_mapping) const
	Creates a shared_ptr clone of the factor with different keys using a map from old->new keys.
virtual shared_ptr	rekey (const KeyVector &new_keys) const
	Clones a factor and fully replaces its keys.
virtual bool	sendable () const
	Should the factor be evaluated in the same thread as the caller This is to enable factors that has shared states (like the Python GIL lock).
Public Member Functions inherited from gtsam::Factor
virtual	~Factor ()=default
	Default destructor.
bool	empty () const
	Whether the factor is empty (involves zero variables).
Key	front () const
	First key.
Key	back () const
	Last key.
const_iterator	find (Key key) const
	find
const KeyVector &	keys () const
	Access the factor's involved variable keys.
const_iterator	begin () const
	Iterator at beginning of involved variable keys.
const_iterator	end () const
	Iterator at end of involved variable keys.
size_t	size () const
virtual void	printKeys (const std::string &s="Factor", const KeyFormatter &formatter=DefaultKeyFormatter) const
	print only keys
bool	equals (const This &other, double tol=1e-9) const
	check equality
KeyVector &	keys ()
iterator	begin ()
	Iterator at beginning of involved variable keys.
iterator	end ()
	Iterator at end of involved variable keys.

Static Public Member Functions
static Matrix2N	BuildWnoaCovariance (double timestep, const VectorN &q_psd_diag)
	Build the continuous-time WNOA discretized process covariance.
static Matrix2N	BuildInverseWnoaCovariance (double dt, const VectorN &q_psd_diag)
	Build the inverse of the WNOA discretized process covariance.
static noiseModel::Gaussian::shared_ptr	BuildWnoaNoiseModel (double timestep, const VectorN &q_psd_diag)
	Convenience helper to construct a Gaussian noise model from q_psd_diag.
static Matrix2N	TransitionFunction (double deltaT)
	Transition matrix for the WNOA prior.
static Matrix2N	ComputeJacobianPrev (const std::pair< Pose, Velocity > &pv1, const std::pair< Pose, Velocity > &pv2, double deltaT)
	Compute interpolation Jacobian with respect to the previous (left) state.
static Matrix2N	ComputeJacobianNext (const std::pair< Pose, Velocity > &pv1, const std::pair< Pose, Velocity > &pv2, double deltaT)
	Compute interpolation Jacobian with respect to the next (right) state.

Static Public Attributes
static constexpr int	dim = traits<Pose>::dimension
Static Public Attributes inherited from gtsam::NoiseModelFactorT< Vector, ValueTypes... >
static constexpr auto	N
	N is the number of variables (N-way factor).

Public Types
using	Velocity = typename gtsam::traits<Pose>::TangentVector
Public Types inherited from gtsam::NoiseModelFactorT< Vector, ValueTypes... >
using	ValueType
	The type of the I'th template param can be obtained as ValueType.
Public Types inherited from gtsam::NoiseModelFactor
typedef std::shared_ptr< This >	shared_ptr
	Noise model.
Public Types inherited from gtsam::NonlinearFactor
typedef std::shared_ptr< This >	shared_ptr
Public Types inherited from gtsam::Factor
typedef KeyVector::iterator	iterator
	Iterator over keys.
typedef KeyVector::const_iterator	const_iterator
	Const iterator over keys.

Additional Inherited Members
Protected Types inherited from gtsam::NoiseModelFactorT< Vector, ValueTypes... >
using	Base
using	This
using	OptionalMatrixTypeT
using	KeyType
using	MatrixTypeT
using	IsConvertible
using	IndexIsValid
using	ContainerElementType
using	IsContainerOfKeys
using	AreAllMatrixRefs
	A helper alias to check if a list of args are all references to a matrix or not.
using	IsMatrixPointer
using	IsNullpointer
using	AreAllMatrixPtrs
	A helper alias to check if a list of args are all pointers to a matrix or not.
Protected Types inherited from gtsam::NoiseModelFactor
typedef NonlinearFactor	Base
typedef NoiseModelFactor	This
Protected Types inherited from gtsam::NonlinearFactor
typedef Factor	Base
typedef NonlinearFactor	This
Protected Member Functions inherited from gtsam::NoiseModelFactor
	NoiseModelFactor (const SharedNoiseModel &noiseModel)
	Constructor - only for subclasses, as this does not set keys.
	Factor ()
	Default constructor for I/O.
template<typename CONTAINER>
	Factor (const CONTAINER &keys)
	Construct factor from container of keys.
template<typename ITERATOR>
	Factor (ITERATOR first, ITERATOR last)
	Construct factor from iterator keys.
template<typename CONTAINER>
static Factor	FromKeys (const CONTAINER &keys)
	Construct factor from container of keys.
template<typename ITERATOR>
static Factor	FromIterators (ITERATOR first, ITERATOR last)
	Construct factor from iterator keys.
Protected Attributes inherited from gtsam::NoiseModelFactor
SharedNoiseModel	noiseModel_
Protected Attributes inherited from gtsam::Factor
KeyVector	keys_
	The keys involved in this factor.

Constructor & Destructor Documentation

◆ WnoaMotionFactor() [1/2]

template<class Pose>

gtsam::WnoaMotionFactor< Pose >::WnoaMotionFactor	(	const StateData &	state_k,
		const StateData &	state_kp1,
		const VectorN &	q_psd_diag )

inline

Construct a WNOA motion factor from two StateData entries.

This constructor builds a factor connecting the pose and velocity keys contained in state_k and state_kp1. The internal noise model is constructed from the provided diagonal PSD vector q_psd_diag and the timestep computed from the two state timestamps.

Parameters

state_k	StateData for time t_k (provides keys and timestamp).
state_kp1	StateData for time t_{k+1} (provides keys and timestamp).
q_psd_diag	Diagonal power spectral density vector used to form the process noise.

◆ WnoaMotionFactor() [2/2]

template<class Pose>

gtsam::WnoaMotionFactor< Pose >::WnoaMotionFactor	(	Key	poseKey0,
		Key	velKey0,
		Key	poseKey1,
		Key	velKey1,
		const double	deltaT,
		const VectorN &	q_psd_diag )

inline

Construct a WNOA factor given explicit keys and timestep.

Parameters

poseKey0	Pose key at t_k
velKey0	Velocity key at t_k
poseKey1	Pose key at t_{k+1}
velKey1	Velocity key at t_{k+1}
deltaT	Time interval t_{k+1} - t_k (must be > 0)
q_psd_diag	Diagonal PSD vector used to form the process noise.

Member Function Documentation

◆ BuildInverseWnoaCovariance()

template<class Pose>

Matrix2N gtsam::WnoaMotionFactor< Pose >::BuildInverseWnoaCovariance	(	double	dt,
		const VectorN &	q_psd_diag )

inlinestatic

Build the inverse of the WNOA discretized process covariance.

Parameters

timestep	Time interval
q_psd_diag	Diagonal PSD vector

Returns: Matrix2N Inverse process covariance matrix

◆ BuildWnoaCovariance()

template<class Pose>

Matrix2N gtsam::WnoaMotionFactor< Pose >::BuildWnoaCovariance	(	double	timestep,
		const VectorN &	q_psd_diag )

inlinestatic

Build the continuous-time WNOA discretized process covariance.

Returns the 2N x 2N covariance matrix for a WNOA prior discretized over timestep using the diagonal PSD q_psd_diag. See (11.7) in (Barfoot, 2024) for the derivation of covariance for the WNOA prior

Parameters

timestep	Time interval over which to compute covariance
q_psd_diag	Diagonal PSD vector

Returns: Matrix2N Process covariance

◆ BuildWnoaNoiseModel()

template<class Pose>

noiseModel::Gaussian::shared_ptr gtsam::WnoaMotionFactor< Pose >::BuildWnoaNoiseModel	(	double	timestep,
		const VectorN &	q_psd_diag )

inlinestatic

Convenience helper to construct a Gaussian noise model from q_psd_diag.

The noise model uses the covariance produced by BuildWnoaCovariance.

Parameters

timestep	Time interval
q_psd_diag	Diagonal PSD vector

Returns: noiseModel::Gaussian::shared_ptr Noise model built from covariance

◆ ComputeJacobianNext()

template<class Pose>

Matrix2N gtsam::WnoaMotionFactor< Pose >::ComputeJacobianNext	(	const std::pair< Pose, Velocity > &	pv1,
		const std::pair< Pose, Velocity > &	pv2,
		double	deltaT )

inlinestatic

Compute interpolation Jacobian with respect to the next (right) state.

Computes the 2N x 2N Jacobian block that maps perturbations in the next bordering state (pose_kp1, vel_kp1) to perturbations in the interpolated discrete residual.

Parameters

pv1	Pair (pose_k, vel_k)
pv2	Pair (pose_kp1, vel_kp1)
deltaT	Time interval

Returns: Matrix2N Jacobian block w.r.t. next state

◆ ComputeJacobianPrev()

template<class Pose>

Matrix2N gtsam::WnoaMotionFactor< Pose >::ComputeJacobianPrev	(	const std::pair< Pose, Velocity > &	pv1,
		const std::pair< Pose, Velocity > &	pv2,
		double	deltaT )

inlinestatic

Compute interpolation Jacobian with respect to the previous (left) state.

Computes the 2N x 2N Jacobian block that maps perturbations in the previous bordering state (pose_k, vel_k) to perturbations in the interpolated discrete residual.

Parameters

pv1	Pair (pose_k, vel_k)
pv2	Pair (pose_kp1, vel_kp1)
deltaT	Time interval

Returns: Matrix2N Jacobian block w.r.t. previous state

◆ equals()

template<class Pose>

bool gtsam::WnoaMotionFactor< Pose >::equals	(	const NonlinearFactor &	expected,
		double	tol = 1e-9 ) const

inlineoverridevirtual

equals

Reimplemented from gtsam::NoiseModelFactor.

◆ evaluateError() [1/5]

template<class Pose>

Vector gtsam::WnoaMotionFactor< Pose >::evaluateError	(	const Pose &	p1,
		const Velocity &	v1,
		const Pose &	p2,
		const Velocity &	v2,
		OptionalMatrixType	Hp1,
		OptionalMatrixType	Hv1,
		OptionalMatrixType	Hp2,
		OptionalMatrixType	Hv2 ) const

inlineoverride

Evaluate the WNOA factor residual and optional Jacobians.

Residual is a 2N vector formed from the difference between the discrete relative pose/velocity and the predicted motion under the WNOA prior. When Jacobian pointers are provided, the function fills the corresponding derivative blocks with respect to the four inputs.

Parameters

p1	Pose at time t_k
v1	Velocity at time t_k
p2	Pose at time t_{k+1}
v2	Velocity at time t_{k+1}
Hp1	Optional output Jacobian w.r.t. p1
Hv1	Optional output Jacobian w.r.t. v1
Hp2	Optional output Jacobian w.r.t. p2
Hv2	Optional output Jacobian w.r.t. v2

Returns: Vector Residual vector of size 2*dim

◆ evaluateError() [2/5]

template<class Pose>

Vector gtsam::NoiseModelFactorT< Vector, ValueTypes >::evaluateError ( const ValueTypes &... x ) const

inline

No-Jacobians requested function overload.

This specializes the version below to avoid recursive calls since this is commonly used.

e.g. const Vector error = factor.evaluateError(pose, point);

◆ evaluateError() [3/5]

template<class Pose>

Vector gtsam::NoiseModelFactorT< Vector, ValueTypes >::evaluateError	(	const ValueTypes &...	x,
		MatrixTypeT< ValueTypes > &...	H ) const

inline

If all the optional arguments are matrices then redirect the call to the one which takes pointers.

To get access to this version of the function from derived classes one will need to use the "using" keyword and specify that like this: public: using NoiseModelFactorT<list the value types here>::evaluateError;

◆ evaluateError() [4/5]

template<class Pose>

AreAllMatrixRefs< Vector, OptionalJacArgs... > gtsam::NoiseModelFactorT< Vector, ValueTypes >::evaluateError	(	const ValueTypes &...	x,
		OptionalJacArgs &&...	H ) const

inline

Some (but not all) optional Jacobians are omitted (function overload) and the jacobians are l-value references to matrices.

e.g. const Vector error = factor.evaluateError(pose, point, Hpose);

◆ evaluateError() [5/5]

template<class Pose>

virtual Vector gtsam::NoiseModelFactorT< Vector, ValueTypes >::evaluateError	(	const ValueTypes &...	x,
		OptionalMatrixTypeT< ValueTypes >...	H ) const

Override evaluateError to finish implementing an n-way factor.

Both the x and H arguments are written here as parameter packs, but when overriding this method, you probably want to explicitly write them out. For example, for a 2-way factor with variable types Pose3 and Point3, you should implement:

Vector evaluateError(
    const Pose3& x1, const Point3& x2,
    OptionalMatrixType H1 = OptionalNone,
    OptionalMatrixType H2 = OptionalNone) const override { ... }

If any of the optional Matrix reference arguments are specified, it should compute both the function evaluation and its derivative(s) in the requested variables.

Parameters

	x	The values of the variables to evaluate the error for. Passed in as separate arguments.
[out]	H	The Jacobian with respect to each variable (optional).

◆ print()

template<class Pose>

void gtsam::WnoaMotionFactor< Pose >::print	(	const std::string &	s = "",
		const KeyFormatter &	keyFormatter = DefaultKeyFormatter ) const

inlineoverridevirtual

print

Reimplemented from gtsam::NoiseModelFactor.

◆ TransitionFunction()

template<class Pose>

Matrix2N gtsam::WnoaMotionFactor< Pose >::TransitionFunction ( double deltaT )

inlinestatic

Transition matrix for the WNOA prior.

Returns the 2N x 2N transition matrix mapping the concatenated state [pose; vel] over interval deltaT using the standard WNOA linearization.

Parameters

deltaT Time interval

Returns: Matrix2N Transition matrix

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

/tmp/gtsam-4.3.0-doxygen.rsXPUS/source/gtsam/nonlinear/WnoaFactor.h

Detailed Description

Public Member Functions

Static Public Member Functions

Static Public Attributes

Public Types

Additional Inherited Members

Constructor & Destructor Documentation

◆ WnoaMotionFactor() [1/2]

◆ WnoaMotionFactor() [2/2]

Member Function Documentation

◆ BuildInverseWnoaCovariance()

◆ BuildWnoaCovariance()

◆ BuildWnoaNoiseModel()

◆ ComputeJacobianNext()

◆ ComputeJacobianPrev()

◆ equals()

◆ evaluateError() [1/5]

◆ evaluateError() [2/5]

◆ evaluateError() [3/5]

◆ evaluateError() [4/5]

◆ evaluateError() [5/5]

◆ print()

◆ TransitionFunction()