/gibbs-rtss

Gibbs-Sampling based Rauch-Tung-Striebel Smoother (Code for ACC 2011 paper)

Primary LanguageMatlab

gibbs-rtss

This software package provides a Matlab implementation of the Gibbs-filter and Gibbs-RTSS as described in

Marc Peter Deisenroth and Henrik Ohlsson "A General Perspective on Gaussian Filtering and Smoothing: Explaining Current and Deriving New Algorithms" in Proceedings of the 2011 American Control Conference (ACC 2011)

The software package also contains implementations of the following filters/smoothers:

  • Gibbs-filter/Gibbs-RTSS
  • EKF/EKS
  • UKF/URTSS
  • CKF/CKS

Running demo_nonlinear_model reproduces the results from the paper for the nonlinear example in the paper (figures and numbers).

The code requires MatlabR2007a or newer.

(C) Copyright 2016 by Marc Deisenroth

Permission is granted for anyone to copy, use, or modify this software and accompanying documents for any uncommercial purposes, provided this copyright notice is retained, and note is made of any changes that have been made. This software and documents are distributed without any warranty.

I'd appreciate any feedback on the code (useful, buggy, inefficient, ...)