EUVIMEDEUVIMED entdecken
Crossref

Nonlinear Bayesian Filtering With N atural Gr a dient Gaussia n Appr o ximation

Wenhan Cao, Tianyi Zhang, Zeju Sun, Chang Liu, Stephen S-T Yau, Shengbo Eben Li

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2026-07 · 48 (7) · 8581-8597

DOI: 10.1109/tpami.2026.3674120 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2026-07. Verfasst von Wenhan Cao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 61 Literaturangaben. Gefördert von: NSF China, Beijing Natural Science Foundation, Tsinghua University-Toyota Joint Research Center for AI Technology of Automated Vehicle.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Practical Bayes filters often assume the state distribution of each time step to be Gaussian for computational tractability, resulting in the so-called Gaussian filters. When facing nonlinear systems, Gaussian filters such as extended Kalman filter (EKF) or unscented Kalman filter (UKF) typically rely on certain linearization techniques, which can introduce large estimation errors. To address this issue, this paper reconstructs the prediction and update steps of Gaussian filtering as solutions to two distinct optimization problems, whose optimal conditions are found to have analytical forms from Stein's lemma. It is observed that the stationary point for the prediction step requires calculating the first two moments of the prior distribution, which is equivalent to that step in existing moment-matching filters. In the update step, instead of linearizing the model to approximate the stationary points, we propose an iterative approach to directly minimize the update step's objective to avoid linearization errors. For the purpose of performing the steepest descent on the Gaussian manifold, we derive its natural gradient that leverages Fisher information matrix to adjust the gradient direction, accounting for the curvature of the parameter space. Combining this update step with moment matching in the prediction step, we introduce a new iterative filter for nonlinear systems called Natural Gradient Gaussian Approximation filter, or NANO filter for short. We prove that NANO filter locally converges to the optimal Gaussian approximation at each time step. Furthermore, the estimation error is proven exponentially bounded for nearly linear measurement equation and low noise levels through constructing a supermartingale-like property across consecutive time steps. Real-world experiments demonstrate that, compared to popular Gaussian filters such as EKF, UKF, iterated EKF, and posterior linearization filter, NANO filter reduces the average root mean square error by approximately 45% while maintaining a comparable computational burden.

Förderung

  • NSF China
  • Beijing Natural Science Foundation
  • Tsinghua University-Toyota Joint Research Center for AI Technology of Automated Vehicle

Zitieren

APA

Cao, W., Zhang, T., Sun, Z., Liu, C., Yau, S. S., & Li, S. E. (2026). Nonlinear Bayesian Filtering With N atural Gr a dient Gaussia n Appr o ximation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 48(7), 8581-8597. https://doi.org/10.1109/tpami.2026.3674120

Vancouver

Cao W, Zhang T, Sun Z, Liu C, Yau SS, Li SE. Nonlinear Bayesian Filtering With N atural Gr a dient Gaussia n Appr o ximation. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2026;48(7):8581-8597. doi:10.1109/tpami.2026.3674120

BibTeX
@article{Cao2026,
  author = {Cao, Wenhan and Zhang, Tianyi and Sun, Zeju and Liu, Chang and Yau, Stephen S-T and Li, Shengbo Eben},
  title = {Nonlinear Bayesian Filtering With N atural Gr a dient Gaussia n Appr o ximation},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2026},
  volume = {48},
  number = {7},
  pages = {8581--8597},
  doi = {10.1109/tpami.2026.3674120}
}
Kostenlos starten →