gtsam::HybridSmoother Class Reference

Public Member Functions
	HybridSmoother (const std::optional< double > marginalThreshold={})
	Constructor.
const DiscreteValues &	fixedValues () const
	Return fixed values:
void	reInitialize (HybridBayesNet &&hybridBayesNet)
	Re-initialize the smoother from a new hybrid Bayes Net.
void	reInitialize (HybridBayesNet &hybridBayesNet)
	Re-initialize the smoother from a new hybrid Bayes Net (non rvalue version).
void	update (const HybridNonlinearFactorGraph &graph, const Values &initial, std::optional< size_t > maxNrLeaves={}, const std::optional< Ordering > givenOrdering={})
	Given new factors, perform an incremental update.
Ordering	getOrdering (const HybridGaussianFactorGraph &factors, const KeySet &newFactorKeys)
	Get an elimination ordering which eliminates continuous and then discrete.
std::pair< HybridGaussianFactorGraph, HybridBayesNet >	addConditionals (const HybridGaussianFactorGraph &graph, const HybridBayesNet &hybridBayesNet) const
	Add conditionals from previous timestep as part of liquefication.
HybridGaussianConditional::shared_ptr	gaussianMixture (size_t index) const
	Get the hybrid Gaussian conditional from the Bayes Net posterior at index.
const HybridBayesNet &	hybridBayesNet () const
	Return the Bayes Net posterior.
HybridValues	optimize () const
	Optimize the hybrid Bayes Net, taking into accound fixed values.
void	relinearize (const std::optional< Ordering > givenOrdering={})
	Relinearize the nonlinear factor graph with the latest stored linearization point.
Values	linearizationPoint () const
	Return the current linearization point.
HybridNonlinearFactorGraph	allFactors () const
	Return all the recorded nonlinear factors.
double	error (const HybridValues &x) const
	Compute the linear error using the underlying solver for the explicitly provided discrete assignment.
double	error (const VectorValues &x) const
	Compute the linear error using the underlying solver.

Constructor & Destructor Documentation

HybridSmoother()

gtsam::HybridSmoother::HybridSmoother ( const std::optional< double > marginalThreshold = {} )

inline

Constructor.

Parameters

removeDeadModes	Flag indicating whether to remove dead modes.
marginalThreshold	The threshold above which a mode gets assigned a value and is considered "dead". 0.99 is a good starting value.

Member Function Documentation

addConditionals()

std::pair< HybridGaussianFactorGraph, HybridBayesNet > gtsam::HybridSmoother::addConditionals	(	const HybridGaussianFactorGraph &	graph,
		const HybridBayesNet &	hybridBayesNet ) const

Add conditionals from previous timestep as part of liquefication.

Parameters

graph	The new factor graph for the current time step.
hybridBayesNet	The hybrid bayes net containing all conditionals so far.
ordering	The elimination ordering.

Returns

std::pair<HybridGaussianFactorGraph, HybridBayesNet>

error() [1/2]

double gtsam::HybridSmoother::error

(

const HybridValues &

x

)

const

Compute the linear error using the underlying solver for the explicitly provided discrete assignment.

Parameters

x	The vector and discrete values to compute the error for.

Returns

double The error value.

error() [2/2]

double gtsam::HybridSmoother::error

(

const VectorValues &

x

)

const

Compute the linear error using the underlying solver.

The error is computed using the computed Most-Probable Explanation (MPE) of the discrete variables.

Parameters

x	The vector values to compute the error for.

Returns

double The error value.

gaussianMixture()

HybridGaussianConditional::shared_ptr gtsam::HybridSmoother::gaussianMixture

(

size_t

index

)

const

Get the hybrid Gaussian conditional from the Bayes Net posterior at index.

Parameters

index

Indexing value.

Returns

HybridGaussianConditional::shared_ptr

getOrdering()

Ordering gtsam::HybridSmoother::getOrdering	(	const HybridGaussianFactorGraph &	factors,
		const KeySet &	newFactorKeys )

Get an elimination ordering which eliminates continuous and then discrete.

Expects factors to already have the necessary conditionals which were connected to the variables in the newly added factors. Those variables should be in newFactorKeys.

Parameters

factors	All the new factors and connected conditionals.
newFactorKeys	The keys/variables in the newly added factors.

Returns

Ordering

relinearize()

void gtsam::HybridSmoother::relinearize

(

const std::optional< Ordering >

givenOrdering = {}

)

Relinearize the nonlinear factor graph with the latest stored linearization point.

Parameters

givenOrdering

An optional elimination ordering.

update()

void gtsam::HybridSmoother::update	(	const HybridNonlinearFactorGraph &	graph,
		const Values &	initial,
		std::optional< size_t >	maxNrLeaves = {},
		const std::optional< Ordering >	givenOrdering = {} )

Given new factors, perform an incremental update.

The relevant densities in the hybridBayesNet will be added to the input graph (fragment), and then eliminated according to the ordering presented. The remaining factor graph contains hybrid Gaussian factors that are not connected to the variables in the ordering, or a single discrete factor on all discrete keys, plus all discrete factors in the original graph.

Note

If maxNrLeaves is given, we look at the discrete factor resulting from this elimination, and prune it and the Gaussian components corresponding to the pruned choices.

Parameters

graph	The new factors, should be linear only
maxNrLeaves	The maximum number of leaves in the new discrete factor, if applicable
givenOrdering	The (optional) ordering for elimination, only continuous variables are allowed

Prune

---

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

• /tmp/gtsam-4.3.0-doxygen.rsXPUS/source/gtsam/hybrid/HybridSmoother.h
• /tmp/gtsam-4.3.0-doxygen.rsXPUS/source/gtsam/hybrid/HybridSmoother.cpp