The unscented Kalman filter for nonlinear estimation

  title={The unscented Kalman filter for nonlinear estimation},
  author={Eric A. Wan and R. Van Der Merwe},
  journal={Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat. No.00EX373)},
  • E. Wan, R. Van Der Merwe
  • Published 1 October 2000
  • Mathematics
  • Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat. No.00EX373)
This paper points out the flaws in using the extended Kalman filter (EKE) and introduces an improvement, the unscented Kalman filter (UKF), proposed by Julier and Uhlman (1997). A central and vital operation performed in the Kalman filter is the propagation of a Gaussian random variable (GRV) through the system dynamics. In the EKF the state distribution is approximated by a GRV, which is then propagated analytically through the first-order linearization of the nonlinear system. This can… 

Figures from this paper

7 the Unscented Kalman Filter
This chapter points out the underlying assumptions and flaws in the EKF, and presents an alternative filter with performance superior to that of the EkF, called the unscented Kalman filter (UKF).
Extended and unscented Kalman filters for the identification of uncertainties in a process
  • M. Nasri, W. Kinsner
  • Engineering
    2013 IEEE 12th International Conference on Cognitive Informatics and Cognitive Computing
  • 2013
Simulation results indicate that the EKF and UKF perform equally well when small initial errors are present, however, when large errors are introduced, the UKF leads to a faster convergence and achieves a higher more accurate estimate of the state of the system.
A Comparison Of Estimation Accuracy By The UseOf KF, EKF & UKF Filters
This paper considers the problem of applying the Kalman filters to nonlinear systems. The Kalman filter (KF) is an optimal linear estimator when the process noise and the measurement noise can be
The square-root unscented Kalman filter for state and parameter-estimation
  • Rudolph van der Merwe, E. Wan
  • Mathematics
    2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221)
  • 2001
The square-root unscented Kalman filter (SR-UKF) is introduced which is also O(L/sup 3/) for general state estimation and O( L/sup 2/) for parameter estimation and has the added benefit of numerical stability and guaranteed positive semi-definiteness of the state covariances.
On Iterative Unscented Kalman Filter using Optimization
Six iterated UKF (IUKF) variations are deriving based on two cost functions and three optimization algorithms, which show that IUKF algorithms can be used as a derivative free alternative to IEKF, and provide insights about the different design choices available in IukF algorithms.
Some Relations Between Extended and Unscented Kalman Filters
It is shown that the sigma point function evaluations can be used in the classical EKF rather than an explicitly linearized model, and a less cited version of the EKf based on a second-order Taylor expansion is shown to be quite closely related to UKF.
Unscented Kalman Filtering of a Simulated pH System
This work compares the performance of the two approaches in a simulated pH process with three situations considered, and finds the unscented Kalman filter produced more-accurate results.
State estimation for a tractor-trailer system using adaptive unscented Kalman filter
In this paper, an adaptive unscented Kalman filter (AUKF) is developed for state estimation on a tractor-trailer system. For nonlinear systems, the unscented Kalman filter (UKF) is a well-known
Unscented Kalman filter with nonlinear dynamic process modeling for GPS navigation
This paper preliminarily investigates the application of unscented Kalman filter (UKF) approach with nonlinear dynamic process modeling for Global positioning system (GPS) navigation processing. Many
Unscented Kalman Inversion
It is demonstrated that unscented Kalman methods also provide an effective tool for derivative-free inversion in the setting of black-box forward models, introducing UKI based methods, of particular value for parameter estimation problems in which the number of parameters is moderate but the forward model is expensive and provided as a black box which is impractical to differentiate.


New extension of the Kalman filter to nonlinear systems
It is argued that the ease of implementation and more accurate estimation features of the new filter recommend its use over the EKF in virtually all applications.
Dual Estimation and the Unscented Transformation
This paper points out the flaws in using the EKF, and proposes an improvement based on a new approach called the unscented transformation (UT) [3], which achieves a substantial performance gain with the same order of computational complexity as that of the standard EKf.
The scaled unscented transformation
  • S. Julier
  • Computer Science
    Proceedings of the 2002 American Control Conference (IEEE Cat. No.CH37301)
  • 2002
A generalisation of the unscented transformation (UT) which allows sigma points to be scaled to an arbitrary dimension, giving the same performance as a second order truncated filter but without the need to calculate any Jacobians or Hessians.
Decoupled extended Kalman filter training of feedforward layered networks
These studies demonstrate that the judicious grouping of weights along with the use of artificial process noise in DEKF result in input-output mapping performance that is comparable to the global extended Kalman algorithm, and is often superior to SBP, while requiring significantly fewer presentations of training data than SBP and less overall training time than either of these procedures.
The Unscented Particle Filter
This paper proposes a new particle filter based on sequential importance sampling that outperforms standard particle filtering and other nonlinear filtering methods very substantially and is in agreement with the theoretical convergence proof for the algorithm.
Sequential Monte Carlo methods for optimisation of neural network models
A novel strategy for training neural networks using sequential Monte Carlo algorithms is discussed and a new hybrid gradient descent/sampling importance resampling algorithm (HySIR) is proposed, which is well suited to applications involving on-line, nonlinear and non-Gaussian signal processing.
Training Multilayer Perceptrons with the Extende Kalman Algorithm
It is shown that training multilayer perceptrons is an identification problem for a nonlinear dynamic system which can be solved using the Extended Kalman Algorithm.
Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper
Vibratory power unit for vibrating conveyers and screens comprising an asynchronous polyphase motor, at least one pair of associated unbalanced masses disposed on the shaft of said motor, with the
Neural dual extended Kalman filtering: applications in speech enhancement and monaural blind signal separation
  • E. Wan, A. T. Nelson
  • Engineering
    Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop
  • 1997
A nonlinear time-domain method called dual extended Kalman filtering (DEKF) is presented for removing nonstationary and colored noise from speech and generalizes the algorithm to perform the blind separation of two speech signals from a single recording.
Adaptive Filter Theory