Jafar
|
Base class for all Gaussian innovations defined in the module rtslam. More...
Base class for all Gaussian innovations defined in the module rtslam.
It implements the trivial innovation model:
which is, after all,
so usual in Kalman filtering.
It also returns the Jacobian matrices:
Derive this class and overload the methods if you need other non-trivial innovation models (useful for line landmarks).
Definition at line 45 of file innovation.hpp.
#include <innovation.hpp>
Public Member Functions | |
Innovation (const size_t _size) | |
Size construction. | |
virtual | ~Innovation () |
mandatory virtual destructor | |
void | invertCov () |
the inverse of the innovation covariance. | |
double | mahalanobis () |
The Mahalanobis distance. | |
Public Attributes | |
jblas::sym_mat | iP_ |
The inverse of the innovation covariances matrix. | |
double | mahalanobis_ |
The Mahalanobis distance from the measurement to the expectation. | |
double | relevance |
The Mahalanobis distance of innovation.x wrt measurement.P. |
jafar::rtslam::Innovation::Innovation | ( | const size_t | _size | ) |
Size construction.
Use this constructor for usual innovations with equal expectation, measurement and innovation sizes.
_size | the innovation size |
Generated on Wed Oct 15 2014 00:37:45 for Jafar by doxygen 1.7.6.1 |