The one step fixed-lag particle smoother as a strategy to improve the prediction step of particle filtering
Sequential Monte Carlo methods have been a major advance in the field of numerical filtering for stochastic dynamical state-space systems with partial and noisy observations. However, these methods still have some weaknesses. One of its main weaknesses concerns the degeneracy of these particle filters due to the impoverishment of the particles. Indeed, during the prediction step of these filters, the particles explore the state space, and if this exploration phase is not done correctly, a large part of the particles will end up in areas that are weakly weighted by the new measurement and will be mostly eliminated. Only a few particles will remain, leading to a degeneracy of the filter. In order to improve this last step within the framework of the classic bootstrap particle filter, we propose a simple approximation of the one step fixed-lag smoother. At each time iteration, we propose to perform additional simulations during the prediction step in order to improve the likelihood of the selected particles. Note that we aim to propose an algorithm that is almost as fast and of the same order of complexity as the bootstrap particle filter, and which is robust in poorly conditioned filtering situations. We also investigate a robust version of this smoother.
| Auteurs principaux: | , , , |
|---|---|
| Format: | article biblioteca |
| Langue: | eng |
| Sujets: | U30 - Méthodes de recherche, U10 - Informatique, mathématiques et statistiques, modèle de simulation, modèle mathématique, application des ordinateurs, http://aims.fao.org/aos/agrovoc/c_24242, http://aims.fao.org/aos/agrovoc/c_24199, http://aims.fao.org/aos/agrovoc/c_24009, |
| Accès en ligne: | http://agritrop.cirad.fr/607596/ http://agritrop.cirad.fr/607596/1/nyobe23.pdf |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|