This is an example Matlab implementation of the Rao-Blackwellized Gaussian filtering and smoothing algorithms introduced in:

• R. Hostettler and S. Särkkä, “Rao–Blackwellized Gaussian smoothing,” IEEE Transactions on Automatic Control, vol. 64, no. 1, pp. 302–309, January 2019.

  [Link] [PDF]

The numerical example is implemented in the file rbgs_example.m, while the folder lib contains the algorithm implementation, helper functions, and implementations of the compared algorithms.

Known issues:

• Iterative variants are not implemented.
• Variants for non-additive noise are not implemented.
• Defaults for the sigma-points in gf, gf_*, rbgf, and rbgf_* are missing.