Institute Homepage
Institute Homepage DE Sign In

Back

Autonomous Motion Intelligent Control Systems Conference Paper 2016

Robust Gaussian Filtering using a Pseudo Measurement

Thumb ticker sm 7028ad2b e106 4b00 8e44 482ac9bd2228
Empirical Inference
Manuel Wüthrich
  • Postdoctoral Researcher
Thumb ticker sm pic
Autonomous Motion, Perceiving Systems
Cristina Garcia Cifuentes
Thumb ticker sm 2018 ac r7b9314 cut
Intelligent Control Systems
Sebastian Trimpe
Thumb ticker sm franzi smallr
Autonomous Motion
Franzi Meier
Thumb ticker sm me portrait
Autonomous Motion
Jeannette Bohg
  • Affiliated Researcher
Thumb ticker sm jan profile
Autonomous Motion
Jan Issac
Thumb ticker sm ss
Autonomous Motion
Stefan Schaal
  • Director
Screen shot 2015 12 04 at 15.11.43

Most widely-used state estimation algorithms, such as the Extended Kalman Filter and the Unscented Kalman Filter, belong to the family of Gaussian Filters (GF). Unfortunately, GFs fail if the measurement process is modelled by a fat-tailed distribution. This is a severe limitation, because thin-tailed measurement models, such as the analytically-convenient and therefore widely-used Gaussian distribution, are sensitive to outliers. In this paper, we show that mapping the measurements into a specific feature space enables any existing GF algorithm to work with fat-tailed measurement models. We find a feature function which is optimal under certain conditions. Simulation results show that the proposed method allows for robust filtering in both linear and nonlinear systems with measurements contaminated by fat-tailed noise.

Author(s): Wüthrich, M. and Garcia Cifuentes, C. and Trimpe, S. and Meier, F. and Bohg, J. and Issac, J. and Schaal, S.
Book Title: Proceedings of the American Control Conference (ACC)
Year: 2016
Month: July
Project(s):
Bibtex Type: Conference Paper (inproceedings)
Address: Boston, MA, USA
DOI: 10.1109/ACC.2016.7525473
State: Published
URL: http://arxiv.org/abs/1509.04072
Electronic Archiving: grant_archive
Links:

BibTex 

@inproceedings{wuthrich-rgf-2015,
  title = {Robust Gaussian Filtering using a Pseudo Measurement},
  booktitle = {Proceedings of the American Control Conference (ACC)},
  abstract = {Most widely-used state estimation algorithms, such as the Extended Kalman Filter and the Unscented Kalman Filter, belong to the family of Gaussian Filters (GF). Unfortunately, GFs fail if the measurement process is modelled by a fat-tailed distribution. This is a severe limitation, because thin-tailed measurement models, such as the analytically-convenient and therefore widely-used Gaussian distribution, are sensitive to outliers. In this paper, we show that mapping the measurements into a specific feature space enables any existing GF algorithm to work with fat-tailed measurement models. We find a feature function which is optimal under certain conditions. Simulation results show that the proposed method allows for robust filtering in both linear and nonlinear systems with measurements contaminated by fat-tailed noise.},
  address = {Boston, MA, USA},
  month = jul,
  year = {2016},
  slug = {wuthrich-rgf-2015},
  author = {W{\"u}thrich, M. and Garcia Cifuentes, C. and Trimpe, S. and Meier, F. and Bohg, J. and Issac, J. and Schaal, S.},
  url = {http://arxiv.org/abs/1509.04072},
  month_numeric = {7}
}