Skip to content

Nonlinear Sigma-Point Kalman Filters based on Bayesian Quadrature

License

Notifications You must be signed in to change notification settings

jacobnzw/SSMToybox

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SSM Toybox

Python 3 implementation of the nonlinear sigma-point Kalman filters based on Bayesian quadrature. The filter is understood to be composed of moment transforms, which lend it it's uniqueness, which is why the "moment transform" is the key term used troughout the publications and the toybox.

Structure

The toybox itself is confined in the ssmtoybox folder together with the unit tests.

Under research is a code for reproducing results in the following publications (chronological order):

  • gpq: Gaussian Process Quadrature Moment Transform [1]
  • gpqd: Gaussian Process Quadrautre Moment Transform with Derivatives [2]
  • tpq: Student's t-Process Quadrature Moment Transform [3]
  • bsq: Bayes-Sard Quadrature Moment Transform [4]

Build documentation

cd docs
sphinx-apidoc -o ./ ../ssmtoybox ../ssmtoybox/tests
make html

Why toybox?

The aim of this project was mainly to provide a code base for testing ideas during my Ph.D. research related to the application of Bayesian quadrature for improvement of Kalman filter estimates in terms of crediblity. The code was never meant to be used seriously as a toolbox.

References

[1]: [DOI | PDF] Prüher, J. & Straka, O. Gaussian Process Quadrature Moment Transform, IEEE Transactions on Automatic Control, 2017

[2]: [DOI | PDF] Prüher, J., & Sarkka, S. (2016). On the Use of Gradient Information in Gaussian Process Quadratures. In 2016 IEEE 26th International Workshop on Machine Learning for Signal Processing (MLSP) (pp. 1–6).

[3]: [DOI | PDF] Prüher, J.; Tronarp, F.; Karvonen, T.; Särkkä, S. & Straka, O. Student-t Process Quadratures for Filtering of Non-linear Systems with Heavy-tailed Noise, 20th International Conference on Information Fusion (Fusion), 1-8, 2017

[4]: [DOI | PDF] Pruher, J., Karvonen, T., Oates, C. J., Straka, O., & Sarkka, S. (2021). Improved Calibration of Numerical Integration Error in Sigma-Point Filters. IEEE Transactions on Automatic Control, 66(3), 1286–1292.