parzen_filter.zip_parzen_resampling_split

Parzen particle filtering as described in Lehn-Schi�ler, T. , Erdogmus, D. and Principe, J. C., "Parzen Particle Filters" , Submitted to ICASSP 2004, and SIR particle filter as described in Arulampalam et. al. "A tutorial on Particle Filters for On-line Non-linear/Non-Gausian Bayesian Tracking". The algorithms is split in a filter step and a resampling step, resampling can be applied when needed, in the main program, resampling is applied after each step. Run "drawerror.m" to get started. The function "generateSystem" is used to set up the system equations and noise models. To run with Gamma noise, a function to sample from the distribution is needed, one can be found in Kevin Murphys Bayes Net Toolbox, availible at http://www.ai.mit.edu/~murphyk/Software/BNT/bnt.html