Title :
The

Fixed-Interval Smoothing Problem for Continuous Systems
Author :
Blanco, Eric ; Neveux, Philippe ; Thomas, Gérard
Author_Institution :
Ecole Centrale de Lyon, Ecully
Abstract :
The Hinfin smoothing problem for continuous systems is treated in a state space representation by means of variational calculus techniques. The smoothing problem is introduced in an Hinfin criterion by means of an artificial discontinuity that splits the problem in term of Hinfin forward and Hinfin backward filtering problems. Hence, the smoother design is realized in three steps. First, a forward filter is developed. Secondly, a backward filter is developed taking into account the backward Markovian model. The third step consists of combining the two previous steps in order to compute the Hinfin smoothed estimate. An example shows the efficiency of this proposed smoother
Keywords :
Hinfin optimisation; Markov processes; smoothing methods; variational techniques; Hinfin backward filtering; Hinfin fixed-interval smoothing problem; backward Markovian model; forward filter; state space representation; variational calculus techniques; Calculus; Continuous time systems; Filtering; Filters; Noise level; Riccati equations; Smoothing methods; State estimation; State-space methods; Statistics; Markovian model; robust estimation; signal processing; smoothing; variational calculus;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2006.881237