Emrys::FilterUpdate< MeasurementSpace, Count, Scalar_, Allocator > Struct Template Reference

What a filter's update did: the branches of the update before reduction, with their weights and innovations but not their posteriors (decisions.md [innovation-statistics]), and whether the reduction policy accepted the measurement. More...

#include <EmrysFilter/Models/MeasurementUpdate.h>

Public Types
using	Scalar = Scalar_
	the scalar type
using	MeasurementSpaceType = MeasurementSpace
	the measurement space
using	BranchType = FilterBranch<MeasurementSpace, Scalar>
	one branch
using	allocator_type = Allocator
	where a run-time number of branches comes from
using	Container
	The branches' container: an array when their number is known at compile time, and otherwise a buffer from Allocator.

Public Member Functions
	FilterUpdate ()=default
	no branches, not accepted
	FilterUpdate (Container branches, Scalar mixtureLogLikelihood=0, bool accepted=false)
	The branches branches, as MeasurementUpdate(branches, mixtureLogLikelihood), and whether the policy accepted the measurement, accepted.
Scalar	logLikelihood () const
Eigen::Index	innovationSize () const
Allocator	get_allocator () const

Public Attributes
Container	branches
	one per noise component, weights summing to one
bool	accepted = false
	whether the policy accepted the measurement

Static Public Attributes
static constexpr std::size_t	BranchCount = Count
	the number of branches, or std::dynamic_extent

Detailed Description

template<typename MeasurementSpace, std::size_t Count = std::dynamic_extent, typename Scalar_ = double, typename Allocator = DefaultAllocator>
struct Emrys::FilterUpdate< MeasurementSpace, Count, Scalar_, Allocator >

What a filter's update did: the branches of the update before reduction, with their weights and innovations but not their posteriors (decisions.md [innovation-statistics]), and whether the reduction policy accepted the measurement.

The log-likelihood and each branch's NIS are computed from the innovations when first asked for, unless the filter had them already. When accepted is false, the policy (such as a Gate) rejected the measurement and the state kept its prior. Pass it to a NisMonitor as any update.

Template Parameters

MeasurementSpace	the measurement space
Count	number of branches if known at compile time, otherwise std::dynamic_extent
Scalar_	the scalar type
Allocator	where a run-time number of branches comes from (see Allocation)

Member Typedef Documentation

Container

template<typename MeasurementSpace, std::size_t Count = std::dynamic_extent, typename Scalar_ = double, typename Allocator = DefaultAllocator>

using Emrys::FilterUpdate< MeasurementSpace, Count, Scalar_, Allocator >::Container

Initial value:

std::conditional_t<
       Count == std::dynamic_extent,
       internal::Buffer<BranchType, Allocator>,
       std::array<BranchType, Count>>

The branches' container: an array when their number is known at compile time, and otherwise a buffer from Allocator.

Member Function Documentation

get_allocator()

template<typename MeasurementSpace, std::size_t Count = std::dynamic_extent, typename Scalar_ = double, typename Allocator = DefaultAllocator>

Allocator Emrys::FilterUpdate< MeasurementSpace, Count, Scalar_, Allocator >::get_allocator ( ) const

inlinenodiscard

Returns

the allocator a run-time number of branches comes from, or a default-constructed one for an array of them

innovationSize()

template<typename MeasurementSpace, std::size_t Count = std::dynamic_extent, typename Scalar_ = double, typename Allocator = DefaultAllocator>

Eigen::Index Emrys::FilterUpdate< MeasurementSpace, Count, Scalar_, Allocator >::innovationSize ( ) const

inlinenodiscard

Returns

the difference dimension of the innovation, the degrees of freedom of each NIS

logLikelihood()

template<typename MeasurementSpace, std::size_t Count = std::dynamic_extent, typename Scalar_ = double, typename Allocator = DefaultAllocator>

Scalar Emrys::FilterUpdate< MeasurementSpace, Count, Scalar_, Allocator >::logLikelihood ( ) const

inlinenodiscard

Returns

the observation's log-likelihood, \( \log p(z) = \log \sum_k w_k \, p(z \mid k) \)

---

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

• EmrysFilter/Models/MeasurementUpdate.h