One outcome of an update: a posterior, how probable it is, and its innovation, from which the observation's likelihood under it and its NIS follow (decisions.md [innovation-statistics]).
More...
#include <EmrysFilter/Models/MeasurementUpdate.h>
|
|
Scalar | logWeight |
| | log of its posterior probability among the branches of one update
|
|
GaussianParams< Space, Scalar > | posterior |
| | the state conditioned on the observation under this branch
|
|
Innovation< MeasurementSpace, Scalar > | innovation |
| | the observation's residual under it and its covariance
|
|
Eigen::Index | mode = 0 |
| | flat index of the mode's discrete value, set by the IMM; update() gives 0
|
|
Eigen::Index | component |
| | index of the noise component, such as inlier or outlier
|
|
std::size_t | iterations = 1 |
| | the linearizations the update made: 1, except under the Iterated method
|
|
bool | converged = true |
| | false when an Iterated update stopped at its iteration cap rather than by its tolerance
|
template<typename Space, typename MeasurementSpace, typename Scalar = double>
struct Emrys::UpdateBranch< Space, MeasurementSpace, Scalar >
One outcome of an update: a posterior, how probable it is, and its innovation, from which the observation's likelihood under it and its NIS follow (decisions.md [innovation-statistics]).
- Template Parameters
-
| Space | the state's space |
| MeasurementSpace | the measurement space |
◆ logLikelihood()
template<typename Space, typename MeasurementSpace, typename Scalar = double>
- Returns
- the log-likelihood of the observation under this branch
◆ nis()
template<typename Space, typename MeasurementSpace, typename Scalar = double>
- Returns
- the normalized innovation squared, chi-squared with the innovation's difference dimension as its degrees of freedom for a consistent filter (see Innovation::nis())
The documentation for this struct was generated from the following file:
1.18.0