An improved SIR-based sequential Monte Carlo algorithm

Roland Lamberti (1, 2, 3) , Yohan Petetin (4, 1, 2, 3) , François Septier (5, 6) , François Desbouvries (4, 1, 2, 3)

1 TSP - CITI - Communications, Images et Traitement de l'Information
2 SAMOVAR - Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux
3 CNRS - Centre National de la Recherche Scientifique
4 TIPIC-SAMOVAR - Traitement de l'Information Pour Images et Communications
5 CRIStAL - Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189
6 Institut TELECOM/TELECOM Lille1

Sequential Monte Carlo (SMC) algorithms are based on importance sampling (IS) techniques. Resampling has been introduced as a tool for fighting the weight degeneracy problem. However, for a fixed sample size N, the resampled particles are dependent, are not drawn exactly from the target distribution, nor are weighted properly. In this paper, we revisit the resampling mechanism and propose a scheme where the resampled particles are (conditionally) independent and weighted properly. We validate our results via simulations.

Domaines: Traitement du signal et de l'image [eess.SP]

Roland Lamberti, Yohan Petetin, François Septier, François Desbouvries. An improved SIR-based sequential Monte Carlo algorithm. 2016 IEEE Workshop on Statistical Signal Processing (SSP 16), Jun 2016, Palma de Mallorca, Spain. ⟨hal-01359095⟩