Title :
Coincident smoothing
Author_Institution :
S.L. Fagan Associates, New York, N.Y., USA
fDate :
10/1/1973 12:00:00 AM
Abstract :
It is shown how the matrix inversion implicit in optimal smoothing can be avoided. This is done by regarding the fact that forward and backward filters are estimating the same thing as an "observation." By employing the Measurement Matrix Partitioning Theorem the inversions are avoided.
Keywords :
Smoothing methods; Covariance matrix; Filtering theory; Filters; Laplace equations; Sea measurements; Smoothing methods;
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.1973.1100403