Resampling methods for particle filtering

Implementation of four resampling methods (Multinomial, Residual, Stratified, and Systematic)

Description

------------------------------------------
       Particle filter resampling

 Author: Jose Luis Blanco
 Web: http://www.isa.uma.es/jlblanco
-------------------------------------------

This MATLAB toolkit contains functions and scripts demonstrating four different approaches to resampling:
- Multinomial
- Residual
- Stratified
- Systematic (universal sampling).

The main entry points for this toolkit are:
- runAnalysis.m
- runExample.m

A C++ version of these algorithms are part of the MRPT library ( http://mrpt.sourceforge.net/ ).

• MATLAB release: MATLAB 7.5 (R2007b)

[walid]

thanks

[Paul Mellor]

I'm trying to understand this... you describe it as a SIS, yet i do not see any importance sampling taking place, as no new samples are drawn at each step (only during the resampling step). i had understood that SIS drew samples each step from an importance distribution - or perhaps i've got that wrong?