Convolution is a general operator for linear systems. Consequently, the convolution‐based batch form is common to linear state estimators. This chapter presents the Optimal Finite Impulse Response filter (OFIR) and Limited Memory Filtering theory in both convolution‐based batch form and iterative form using recursions. It also shows OFIR filter is the most general of the other available linear optimal state estimators and that the optimal Kalman recursions serve equivalently to the OFIR filter on finite and infinite horizons. Another useful linear approach to state estimation is to use the likelihood function and obtain an Maximum Likelihood Finite Impulse Response estimate. Extended OFIR filtering is commonly organized by converting a nonlinear state space model to a linear one using the Taylor series and then modifying the linear algorithms. The concepts of a priori filtering and a posteriori filtering play an important role in discrete time.
Optimal FIR and Limited Memory Filtering
Optimal and Robust State Estimation ; 117-157
09.08.2022
41 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Limited memory optimal filtering.
NTRS | 1968
|Suboptimal linear filtering with a limited state
AIAA | 1971
|Desensitized Optimal Filtering
British Library Conference Proceedings | 2011
|Desensitized Optimal Filtering
AIAA | 2011
|