Desensitised Kalman filtering

Karlgaard, Christopher D.; Shen, Honglie

doi:10.1049/iet-rsn.2012.0075

Cited by 35 publications

(27 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…6 shows the mean costs of the three filters. Here the cost of the perfect UKF is computed according to (14), and the cost of the imperfect UKF and the DUKF are computed according to (19). As expected, the perfect UKF yields the smallest cost function, whereas the DUKF yields smaller cost than the imperfect UKF.…”

Section: Linear Examplementioning

confidence: 87%

“…As expected, the perfect UKF yields the smallest cost function, whereas the DUKF yields smaller cost than the imperfect UKF. This demonstrates the effectiveness of the DUKF compared to the imperfect UKF because the DUKF minimises the augmented cost function in (19), whereas the imperfect UKF does not. The comparisons of the RMS errors of the state estimates are presented in Fig.…”

Section: Linear Examplementioning

confidence: 89%

“…The sensitivity penalty is added in (19) regardless of the true α values. As a result, even if by chance the true α values are used in the filter model, the DUKF does not reduce to the standard UKF.…”

Section: Remarkmentioning

confidence: 99%

“…Unlike the DKF and DEKF in [18,19], where evaluation of these quantities is straightforward, it is a rather complicated task to compute s i k and ∂ y k /∂a i because of the manner in which the states are propagated during the unscented transformation. This section presents how these two quantities are computed, which effectively completes the derivation of the DUKF.…”

Section: Propagation Of Sensitivitiesmentioning

confidence: 99%

See 3 more Smart Citations

Sensitivity reduction of unscented Kalman filter about parameter uncertainties

Shen

Karlgaard

2015

IET Radar, Sonar & Navigation

Self Cite

View full text Add to dashboard Cite

This study aims to reduce the sensitivity of the unscented Kalman filter (UKF) estimates with respect to uncertain model parameters, leading to a more robust UKF. The standard minimum-variance cost index is augmented to include a penalty on the sensitivities of the state estimates about parameter uncertainties in the form of a weighted norm. A new filter gain is thus obtained, and the desensitised UKF (DUKF) provides better estimation of the true states than the standard UKF with an imperfect plant model. Numerical examples are shown to demonstrate the efficacy of the DUKF.

show abstract

Section: Linear Examplementioning

confidence: 87%

Section: Linear Examplementioning

confidence: 89%

Section: Remarkmentioning

confidence: 99%

Section: Propagation Of Sensitivitiesmentioning

confidence: 99%

See 2 more Smart Citations

Sensitivity reduction of unscented Kalman filter about parameter uncertainties

Shen

Karlgaard

2015

IET Radar, Sonar & Navigation

Self Cite

View full text Add to dashboard Cite

show abstract

“…So, the local observability of the augmented system cannot be satisfied sometime and the stability of the filter cannot be ensured. The Desensitised Kalman Filter(DKF) [6] is a robust filtering algorithm that is based on the principles of desensitized optimal control. The DKF does not require any structures to be specified or any extra state dimensions, and it is a promising robust estimator.…”

Section: Introductionmentioning

confidence: 99%

Continuous-discrete robust UKF for nonlinear systems with parameter uncertainties

Ishihara

Yamakita²

2017

Stochastic Systems Theory and its Applications (SSS)

View full text Add to dashboard Cite

This paper addresses continuous-discrete filtering problems for parametric uncertain nonlinear systems. We developed a Robust Unsecnted Kalman Filter(RUKF) for discrete-time nonlinear systems with parameter uncertainties in a previous work. The RUKF is more effective than the conventional UKF for uncertain systems. However, the previous RUKF is a discrete time filtering algorithm and it cannot be directly applied to a continuous-discrete filtering problem. So, we modify the predictive step of the RUKF in order to deal with the continuous-discrete filtering problem. The validities of the proposed methods are illustrated in Monte Carlo simulations.

show abstract

Fault Detection for Buildings Using Uncertain Parameters and Interacting Multiple-Model Method

Tabaček,

Havlena

2022

Lecture Notes in Control and Information Sciences - Proceedings

View full text Add to dashboard Cite

Desensitised Kalman filtering

Cited by 35 publications

References 25 publications

Sensitivity reduction of unscented Kalman filter about parameter uncertainties

Sensitivity reduction of unscented Kalman filter about parameter uncertainties

Continuous-discrete robust UKF for nonlinear systems with parameter uncertainties

Fault Detection for Buildings Using Uncertain Parameters and Interacting Multiple-Model Method

Contact Info

Product

Resources

About