KLD-sampling/resampling PF

Attached files are the Matlab implementation of the KLD-sampling/resampling and SIR-based PFs for a bearing-only tracking example

To note, the performance of this tracking example is unstable which can easily unveil the helpless of the particle filter(even the SIR)! In this example, roughening can be very powerful for avoiding sample degeneracy and impoverishment! Please be reminded!!

This may exposes the weakness of filter: a modelling-based estimator suffers greatly from modelling error; see my O2 work to debate this

Since quite few explanations and notes are made in thses .m files, please refer to the following paper for details

T. Li, S. Sun, T. P. Sattar, Adapting sample size in particle filters through KLD-resampling, Electronics Letters, 2013, vol.46, no.12, pp.740-742.

Response to the paper and these codes is welcome and will be appreciated: tiancheng.li1985 at gmail.com; t.c.li at mail.nwpu.edu.cn