Closed-form Filtering for Non-linear Systems - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2024

Closed-form Filtering for Non-linear Systems

Résumé

Sequential Bayesian Filtering aims to estimate the current state distribution of a Hidden Markov Model, given the past observations. The problem is well-known to be intractable for most application domains, except in notable cases such as the tabular setting or for linear dynamical systems with gaussian noise. In this work, we propose a new class of filters based on Gaussian PSD Models, which offer several advantages in terms of density approximation and computational efficiency. We show that filtering can be efficiently performed in closed form when transitions and observations are Gaussian PSD Models. When the transition and observations are approximated by Gaussian PSD Models, we show that our proposed estimator enjoys strong theoretical guarantees, with estimation error that depends on the quality of the approximation and is adaptive to the regularity of the transition probabilities. In particular, we identify regimes in which our proposed filter attains a TV ϵ-error with memory and computational complexity of O(ϵ-1) and O(ϵ-3/2) respectively, including the offline learning step, in contrast to the O(ϵ-2) complexity of sampling methods such as particle filtering.
Fichier principal
Vignette du fichier
CANTELOBRE_PSD_FILTERING.pdf (580.01 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Licence : CC BY - Paternité

Dates et versions

hal-04458936 , version 1 (15-02-2024)

Licence

Paternité

Identifiants

  • HAL Id : hal-04458936 , version 1

Citer

Théophile Cantelobre, Carlo Ciliberto, Benjamin Guedj, Alessandro Rudi. Closed-form Filtering for Non-linear Systems. 2024. ⟨hal-04458936⟩
15 Consultations
11 Téléchargements

Partager

Gmail Facebook X LinkedIn More