unscented Kalman filter

E719011

Bayesian filter nonlinear state estimation algorithm recursive estimator

The unscented Kalman filter is a nonlinear state estimation algorithm that uses a deterministic sampling approach (sigma points) to more accurately capture the mean and covariance of a system than the standard extended Kalman filter.

Try in SPARQL Jump to: Surface forms Disambiguation Statements Elicitation Referenced by

All labels observed (3)

Label	Occurrences
unscented Kalman filter canonical	2
central difference Kalman filter	1
scaled unscented Kalman filter	1

How this entity was disambiguated

This entity first appeared as the object of triple T8216755 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.

NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07

Target entity: unscented Kalman filter
Context triple: [Kalman filter, hasVariant, unscented Kalman filter]

A. Kalman filter
The Kalman filter is a mathematical algorithm used to estimate the changing state of a system from noisy measurements, widely applied in control systems, navigation, and signal processing.
B. extended Kalman filter
The extended Kalman filter is a state estimation algorithm that generalizes the Kalman filter to nonlinear systems by linearizing about the current estimate, widely used in robotics and control for tracking and localization.
C. “A New Approach to Linear Filtering and Prediction Problems”
“A New Approach to Linear Filtering and Prediction Problems” is Rudolf E. Kálmán’s landmark 1960 paper that introduced the Kalman filter, a foundational algorithm for optimal estimation in control theory, signal processing, and navigation.
D. Wiener filter
The Wiener filter is a signal processing technique that optimally estimates a desired signal from noisy observations by minimizing the mean square error, based on statistical properties of signal and noise.
E. Markov localization
Markov localization is a probabilistic method in robotics for estimating a robot’s position by maintaining and updating a belief distribution over all possible locations based on sensor data and motion.
F. None of above. chosen
G. Unsure - the case is ambiguous/there is not enough information to decide.

NED2 Entity disambiguation (via description) gpt-5-mini-2025-08-07

Target entity: unscented Kalman filter
Target entity description: The unscented Kalman filter is a nonlinear state estimation algorithm that uses a deterministic sampling approach (sigma points) to more accurately capture the mean and covariance of a system than the standard extended Kalman filter.

A. Kalman filter
The Kalman filter is a mathematical algorithm used to estimate the changing state of a system from noisy measurements, widely applied in control systems, navigation, and signal processing.
B. extended Kalman filter
The extended Kalman filter is a state estimation algorithm that generalizes the Kalman filter to nonlinear systems by linearizing about the current estimate, widely used in robotics and control for tracking and localization.
C. “A New Approach to Linear Filtering and Prediction Problems”
“A New Approach to Linear Filtering and Prediction Problems” is Rudolf E. Kálmán’s landmark 1960 paper that introduced the Kalman filter, a foundational algorithm for optimal estimation in control theory, signal processing, and navigation.
D. Wiener filter
The Wiener filter is a signal processing technique that optimally estimates a desired signal from noisy observations by minimizing the mean square error, based on statistical properties of signal and noise.
E. Markov localization
Markov localization is a probabilistic method in robotics for estimating a robot’s position by maintaining and updating a belief distribution over all possible locations based on sensor data and motion.
F. None of above. chosen

Statements (49)

Predicate	Object
instanceOf	Bayesian filter ⓘ nonlinear state estimation algorithm ⓘ recursive estimator ⓘ
advantage	higher-order accuracy for nonlinear transformations of Gaussian variables ⓘ no need for linearization ⓘ numerical robustness compared to Jacobian-based methods ⓘ
aimsTo	more accurately capture mean and covariance than extended Kalman filter ⓘ
approximates	posterior covariance ⓘ posterior mean ⓘ
assumes	Gaussian noise ⓘ Gaussian state distribution ⓘ
basedOn	unscented transform ⓘ
comparedTo	extended Kalman filter ⓘ
coreStep	measurement update ⓘ sigma point generation ⓘ time update ⓘ
doesNotRequire	explicit Jacobian computation ⓘ
estimates	state covariance ⓘ state mean ⓘ state of a dynamic system ⓘ
field	control theory ⓘ estimation theory ⓘ signal processing ⓘ
handles	nonlinear measurement models ⓘ nonlinear process models ⓘ
hasVariant	central difference Kalman filter ⓘ linked to: unscented Kalman filter scaled unscented Kalman filter ⓘ linked to: unscented Kalman filter square-root unscented Kalman filter ⓘ
introducedBy	Jeffrey K. Uhlmann ⓘ Simon J. Julier ⓘ
introducedIn	1990s ⓘ
limitation	assumes approximate Gaussianity of distributions ⓘ computational cost grows with state dimension ⓘ
parameter	alpha ⓘ beta ⓘ kappa ⓘ
propagates	sigma points through nonlinear functions ⓘ
relatedTo	Kalman filter ⓘ extended Kalman filter ⓘ particle filter ⓘ
usedIn	aerospace guidance and control ⓘ attitude estimation ⓘ autonomous vehicles ⓘ navigation ⓘ robotics ⓘ sensor fusion ⓘ target tracking ⓘ
uses	deterministic sampling ⓘ sigma points ⓘ

How these facts were elicited

Referenced by (4)

Full triples — surface form annotated when it differs from this entity's canonical label.

Kalman filter → hasVariant → unscented Kalman filter ⓘ

extended Kalman filter → comparedTo → unscented Kalman filter ⓘ

unscented Kalman filter → hasVariant → scaled unscented Kalman filter ⓘ

linked to: unscented Kalman filter

unscented Kalman filter → hasVariant → central difference Kalman filter ⓘ