My Project
Public Member Functions | List of all members
UnscentedObservation Class Referenceabstract

Observation model for an unscented kalman filter. More...

#include <UnscentedKalman.h>

Public Member Functions

virtual void h (cv::Mat &z, cv::Mat &state)=0
 observation model: z = h(state) More...
 
virtual cv::Mat getObservation ()=0
 Returns the current measurement vector. More...
 
virtual cv::Mat getObservationNoise ()=0
 Returns the observation noise covariance matrix. More...
 

Detailed Description

Observation model for an unscented kalman filter.

The implementation needs to allocate correct size measurement vector and noise matrix and to implement a transformation from process state into a measurement.

Definition at line 262 of file UnscentedKalman.h.

Member Function Documentation

◆ getObservation()

virtual cv::Mat getObservation ( )
pure virtual

Returns the current measurement vector.

The returned vector should contain the latest measurement values. In the UKF update phase the process state will be modified in such a way to make the difference between estimated measurement (from method h) and the returned real measurement smaller.

Returns
obs_n size vector containing the current measured values.

◆ getObservationNoise()

virtual cv::Mat getObservationNoise ( )
pure virtual

Returns the observation noise covariance matrix.

The returned matrix will be added to the current observation covariance matrix increasing the uncertainty of measurements. The matrix should reflect the amount of noise in the measurement vector returned by getObservation method.

Returns
obs_n by obs_b matrix containing observation noise covariance; or NULL for no additional noise.

◆ h()

virtual void h ( cv::Mat &  z,
cv::Mat &  state 
)
pure virtual

observation model: z = h(state)

Computes an estimated measurement vector from the current state estimate.

Parameters
zobs_n size vector; The estimated measurement.
statestate_n size vector; The current state.

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