Title :
Discrete-time Smoothing Formulas
Author :
Einicke, Garry A.
Author_Institution :
CSIRO - Exploration & Min., Pullenvale, QLD
Abstract :
The paper presents the discrete-time minimum- variance fixed-interval smoother for linear output estimation problems and an extended Kalman smoother for nonlinear problems. The solutions involve a cascade of a Kalman predictor and an adjoint Kalman predictor. Sufficient conditions for the asymptotic convergence of the underlying Riccati equation and optimal smoother performance are developed. Speech enhancement and nonlinear demodulation examples are described which demonstrate that linear and nonlinear smoothers can provide performance benefits.
Keywords :
Kalman filters; Riccati equations; smoothing methods; Kalman predictor; Kalman smoother; Riccati equation; discrete-time smoothing formula; linear output estimation; minimum-variance fixed-interval smoother; nonlinear problem; Estimation error; Filtering; Frequency domain analysis; Kalman filters; Mercury (metals); Nonlinear equations; Riccati equations; Smoothing methods; Speech enhancement; Sufficient conditions; Discrete-time smoothing; fixed-interval smoothing; non-causal filtering;
Conference_Titel :
TENCON 2005 2005 IEEE Region 10
Conference_Location :
Melbourne, Qld.
Print_ISBN :
0-7803-9311-2
Electronic_ISBN :
0-7803-9312-0
DOI :
10.1109/TENCON.2005.301074
