Obálka

Robust error-state Kalman-type filters for attitude estimation

Uloženo v:
Vydáno v:EURASIP Journal on Advances in Signal Processing vol. 2024, no. 1 (Dec 2024), p. 75
Hlavní autor: Bellés, Andrea
Další autoři: Medina, Daniel, Chauchat, Paul, Labsir, Samy, Vilà-Valls, Jordi
Vydáno:
Springer Nature B.V.
Témata:
State estimation
Data analysis
System dynamics
Robustness (mathematics)
Quaternions
Attitudes
Kalman filters
Satellites
Signal processing
On-line přístup:Citation/Abstract
Full Text - PDF
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstrakt:State estimation techniques appear in a plethora of engineering fields, in particular for the attitude estimation application of interest in this contribution. A number of filters have been devised for this problem, in particular Kalman-type ones, but in their standard form they are known to be fragile against outliers. In this work, we focus on error-state filters, designed for states living on a manifold, here unit-norm quaternions. We propose extensions based on robust statistics, leading to two robust M-type filters able to tackle outliers either in the measurements, in the system dynamics or in both cases. The performance and robustness of these filters is explored in a numerical experiment. We first assess the outlier ratio that they manage to mitigate, and second the type of dynamics outliers that they can detect, showing that the filter performance depends on the measurements’ properties.
ISSN:1687-6172
1687-6180
DOI:10.1186/s13634-024-01172-w
Zdroj:Advanced Technologies & Aerospace Database