tfSmoother: Two-filter Smoother

In eddelbuettel/rcppkalman: 'RcppArmadillo'-Based Kalman Filtering

Description

This function computes the ‘Two filter-based’ Smoother

Usage

tfSmoother(M, P, Y, A, Q, H, R, useinf)

Arguments

M	An N x K matrix of K mean estimates from Kalman filter
P	An N x N x K matrix of K state covariances from Kalman Filter
Y	A D x K matrix of K measurement sequences
A	A N x N state transition matrix.
Q	A N x N process noise covariance matrix.
H	A D x N measurement matrix.
R	A D x D measurement noise covariance.
useinf	An optional boolean variable indicating if information filter should be used (with default true).

Details

This function implements the two filter linear smoother which calculates a “smoothed” sequence from the given Kalman filter output sequence by conditioning all steps to all measurements.

Value

A list with two elements

M

the smoothed state mean sequence, and

P

the smoothes state covariance sequence.

Author(s)

The EKF/UKF Toolbox was written by Simo Särkkä, Jouni Hartikainen, and Arno Solin.

Dirk Eddelbuettel is porting this package to R and C++, and maintaing it.

See Also

kfPredict, kfUpdate, and the documentation for the EKF/UKF toolbox at http://becs.aalto.fi/en/research/bayes/ekfukf

eddelbuettel/rcppkalman documentation built on June 19, 2020, 4:28 a.m.