Tampere University of Technology

TUTCRIS Research Portal

Partitioned Update Kalman Filter

Research output: Contribution to journalArticleScientificpeer-review

Standard

Partitioned Update Kalman Filter. / Raitoharju, Matti; Piché, Robert; Ala-Luhtala, Juha; Ali-Löytty, Simo.

In: Journal of Advances in Information Fusion, Vol. 11, No. 1, 06.2016, p. 3-14.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

Raitoharju, M, Piché, R, Ala-Luhtala, J & Ali-Löytty, S 2016, 'Partitioned Update Kalman Filter', Journal of Advances in Information Fusion, vol. 11, no. 1, pp. 3-14.

APA

Raitoharju, M., Piché, R., Ala-Luhtala, J., & Ali-Löytty, S. (2016). Partitioned Update Kalman Filter. Journal of Advances in Information Fusion, 11(1), 3-14.

Vancouver

Raitoharju M, Piché R, Ala-Luhtala J, Ali-Löytty S. Partitioned Update Kalman Filter. Journal of Advances in Information Fusion. 2016 Jun;11(1):3-14.

Author

Raitoharju, Matti ; Piché, Robert ; Ala-Luhtala, Juha ; Ali-Löytty, Simo. / Partitioned Update Kalman Filter. In: Journal of Advances in Information Fusion. 2016 ; Vol. 11, No. 1. pp. 3-14.

Bibtex - Download

@article{c69d32896211469c93030c14c43ea9d1,
title = "Partitioned Update Kalman Filter",
abstract = "In this paper we present a new Kalman filter extension for state update called Partitioned Update Kalman Filter (PUKF). PUKF updates state using multidimensional measurements in parts. PUKF evaluates the nonlinearity of the measurement function within Gaussian prior by comparing the innovation covariance caused by the second order linearization to the Gaussian measurement noise. A linear transformation is applied to measurements to minimize the nonlinearity of a part of the measurement. The measurement update is applied then using only the part of the measurement that has low nonlinearity and the process is then repeated for the updated state using the remaining part of the transformed measurement until the whole measurement has been used. PUKF does the linearizations numerically and no analytical differentiation is required. Results show that when measurement geometry allows effective partitioning, the proposed algorithm improves estimation accuracy and produces accurate covariance estimates.",
keywords = "math.OC, math.PR",
author = "Matti Raitoharju and Robert Pich{\'e} and Juha Ala-Luhtala and Simo Ali-L{\"o}ytty",
note = "ORG=ase,0.75 ORG=mat,0.25",
year = "2016",
month = "6",
language = "English",
volume = "11",
pages = "3--14",
journal = "Journal of Advances in Information Fusion",
issn = "1557-6418",
publisher = "Information Society of Information Fusion",
number = "1",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Partitioned Update Kalman Filter

AU - Raitoharju, Matti

AU - Piché, Robert

AU - Ala-Luhtala, Juha

AU - Ali-Löytty, Simo

N1 - ORG=ase,0.75 ORG=mat,0.25

PY - 2016/6

Y1 - 2016/6

N2 - In this paper we present a new Kalman filter extension for state update called Partitioned Update Kalman Filter (PUKF). PUKF updates state using multidimensional measurements in parts. PUKF evaluates the nonlinearity of the measurement function within Gaussian prior by comparing the innovation covariance caused by the second order linearization to the Gaussian measurement noise. A linear transformation is applied to measurements to minimize the nonlinearity of a part of the measurement. The measurement update is applied then using only the part of the measurement that has low nonlinearity and the process is then repeated for the updated state using the remaining part of the transformed measurement until the whole measurement has been used. PUKF does the linearizations numerically and no analytical differentiation is required. Results show that when measurement geometry allows effective partitioning, the proposed algorithm improves estimation accuracy and produces accurate covariance estimates.

AB - In this paper we present a new Kalman filter extension for state update called Partitioned Update Kalman Filter (PUKF). PUKF updates state using multidimensional measurements in parts. PUKF evaluates the nonlinearity of the measurement function within Gaussian prior by comparing the innovation covariance caused by the second order linearization to the Gaussian measurement noise. A linear transformation is applied to measurements to minimize the nonlinearity of a part of the measurement. The measurement update is applied then using only the part of the measurement that has low nonlinearity and the process is then repeated for the updated state using the remaining part of the transformed measurement until the whole measurement has been used. PUKF does the linearizations numerically and no analytical differentiation is required. Results show that when measurement geometry allows effective partitioning, the proposed algorithm improves estimation accuracy and produces accurate covariance estimates.

KW - math.OC

KW - math.PR

M3 - Article

VL - 11

SP - 3

EP - 14

JO - Journal of Advances in Information Fusion

JF - Journal of Advances in Information Fusion

SN - 1557-6418

IS - 1

ER -