Trait adskalman::ObservationModel

pub trait ObservationModel<R, SS, OS>
where
    R: RealField,
    SS: DimName,
    OS: DimName + DimMin<OS, Output = OS>,
    DefaultAllocator: Allocator<R, SS, SS> + Allocator<R, SS> + Allocator<R, OS, SS> + Allocator<R, SS, OS> + Allocator<R, OS, OS> + Allocator<R, OS> + Allocator<(usize, usize), OS>,
{
    // Required methods
    fn H(&self) -> &OMatrix<R, OS, SS>;
    fn HT(&self) -> &OMatrix<R, SS, OS>;
    fn R(&self) -> &OMatrix<R, OS, OS>;

    // Provided methods
    fn predict_observation(&self, state: &OVector<R, SS>) -> OVector<R, OS> { ... }
    fn update(
        &self,
        prior: &StateAndCovariance<R, SS>,
        observation: &OVector<R, OS>,
        covariance_method: CovarianceUpdateMethod
    ) -> Result<StateAndCovariance<R, SS>, Error> { ... }
    fn observation_matrix(&self) -> &OMatrix<R, OS, SS> { ... }
    fn observation_matrix_transpose(&self) -> &OMatrix<R, SS, OS> { ... }
    fn observation_noise_covariance(&self) -> &OMatrix<R, OS, OS> { ... }
    fn evaluate(&self, state: &OVector<R, SS>) -> OVector<R, OS> { ... }
}

An observation model, potentially non-linear.

To use a non-linear observation model, the non-linear model must be linearized (e.g. using the prior state estimate) and use this linearization as the basis for a ObservationModel implementation. This would be done every timestep. For an example, see nonlinear_observation.rs.

Required Methods

fn H(&self) -> &OMatrix<R, OS, SS>

Get the observation matrix, H.

fn HT(&self) -> &OMatrix<R, SS, OS>

Get the transpose of the observation matrix, HT.

fn R(&self) -> &OMatrix<R, OS, OS>

Get the observation noise covariance, R.

Provided Methods

fn predict_observation(&self, state: &OVector<R, SS>) -> OVector<R, OS>

For a given state, predict the observation.

The default implementation implements a linear observation model, namely y = Hx where y is the predicted observation, H is the observation matrix, and x is the state. For a non-linear observation model, any implementation of this trait should provide an implementation of this method.

If an observation is not possible, this returns NaN values. (This happens, for example, when a non-linear observation model implements this trait and must be evaluated for a state for which no observation is possible.) Observations with NaN values are treated as missing observations.

fn update( &self, prior: &StateAndCovariance<R, SS>, observation: &OVector<R, OS>, covariance_method: CovarianceUpdateMethod ) -> Result<StateAndCovariance<R, SS>, Error>

Given prior state and observation, estimate the posterior state.

This is the update step in the Kalman filter literature.

fn observation_matrix(&self) -> &OMatrix<R, OS, SS>

👎Deprecated since 0.8.0: Please use the H function instead

Get the observation matrix, H.

fn observation_matrix_transpose(&self) -> &OMatrix<R, SS, OS>

👎Deprecated since 0.8.0: Please use the HT function instead

Get the transpose of the observation matrix, HT.

fn observation_noise_covariance(&self) -> &OMatrix<R, OS, OS>

👎Deprecated since 0.8.0: Please use the R function instead

Get the observation noise covariance, R.

fn evaluate(&self, state: &OVector<R, SS>) -> OVector<R, OS>

👎Deprecated since 0.8.0: Please use the predict_observation function instead

For a given state, predict the observation.

If an observation is not possible, this returns NaN values. (This happens, for example, when a non-linear observation model implements this trait and must be evaluated for a state for which no observation is possible.) Observations with NaN values are treated as missing observations.

Implementors