Robust state estimation and model validation for discrete-time uncertain systems with a deterministic description of noise and uncertainty

doi:10.1016/S0005-1098(97)00188-X

Automatica

Volume 34, Issue 2, February 1998, Pages 271-274

https://doi.org/10.1016/S0005-1098(97)00188-X Get rights and content

Abstract

The paper presents a new approach to robust state estimation for a class of uncertain discrete-time systems with a deterministic description of noise and uncertainty. The main result is a recursive scheme for constructing an ellipsoidal state estimation set of all states consistent with the measured output and the given noise and uncertainty description. The paper also includes a result on model validation whereby it can be determined if the assumed model is consistent with measured data.

References (8)

B. Anderson et al.
D.P. Bertsekas et al.
Recursive state estimation for a set-membership description of uncertainty
IEEE Trans. Automat. Control
(1971)
Y. Chen et al.
Minimax robust deconvolution filters under stochastic parametric and noise uncertainties
IEEE Trans. Signal Process
(1994)
D.J. Clements et al.

There are more references available in the full text version of this article.

Cited by (121)

Improved state estimator for linear-Gaussian systems subject to initialization errors
2022, Chemometrics and Intelligent Laboratory Systems
Citation Excerpt :
Therefore, the initial distributions applied to the estimators are commonly assumed to be identical to each batch, resulting a transient lasting for a while [6] if the KF is employed, as shown in Fig. 1. Various efforts have been devoted to enhancing the robustness against the uncertainties [7,8]. For example, H∞ filter attempts to minimize the worst-case estimated cost, i.e., making the estimated error bound a pre-selected value [9].
This paper proposes an improved state estimator for linear-Gaussian systems subject to initialization errors. A one-step prediction function depicted by the Student-t is reconstructed artificially by inserting an auxiliary variable into the original Gaussian distribution. The variational Bayesian (VB) technique is then employed to obtain the approximated posterior joint distribution of the additional variable and the state. A fixed-point iteration is used for recursions to calculate the necessary moments of state with the updated distribution of the auxiliary variable. Simulation and experiment verify the performance of the proposed algorithm. It shows that the proposed method yields significant improvements over the existing initialization approaches, such as the practical Kalman filter (KF) and the recently-developed Bayesian initialization algorithm.
Distributed set-membership observers for interconnected multi-rate systems
2017, Automatica
Citation Excerpt :
This method relies on bounded uncertainties/disturbances and leads to estimators that provide, in real time, sets containing the state of the system with guarantees. To characterize these sets, different authors have resorted to variants of ellipses (Durieu, Walter, & Polyak, 2001; El Ghaoui & Calafiore, 2001; Savkin & Petersen, 1998), polyhedrons (Kuntsevich & Lychak, 1985), consistency techniques (Jaulin, 2002), interval analysis (Mazenc & Bernard, 2011; Raïssi, Ramdani, & Candau, 2004), or zonotopes (Alamo, Bravo, & Camacho, 2005; Combastel, 2015). The latter approach, derived from parallelotopic descriptions (Chisci, Garulli, & Zappa, 1996), is very suitable for distributed implementations.
This paper presents a new distributed observer for interconnected multi-rate systems. The developed observer belongs to the family of set-membership estimators, and the use of zonotopes is proposed to mathematically describe the sets, a choice motivated by the available mathematical background in operations such as intersections, combinations and linear manipulations. The main features of the proposed distributed observer are (a) the actual state of any subsystem always belongs to the computed sets; (b) the volume of these sets is minimized in real time; (c) under equivalent assumptions, the performance of the observer approaches to that of an analogous distributed Kalman filter.
Gaussian filters for parameter and state estimation: A general review of theory and recent trends
2017, Signal Processing
Citation Excerpt :
Note that the centers of ellipsoids are assumed to be the estimated states. In this context, there are several recursive algorithms to account for uncertain models, particularly the one proposed by Savkin et al. [102]. The Guaranteed-cost design is another important approach in which the filter is designed by preserving an upper bound on the variance of the state estimation error.
Real-time control systems rely on reliable estimates of states and parameters in order to provide accurate and safe control of electro-mechanical systems. The task of extracting state and parametric values from system’s partial measurements is referred to as state and parameter estimation. The main goal is minimizing the estimation error as well as maintaining robustness against the noise and modeling uncertainties. The development of estimation techniques spans over five centuries, and involves a large number of contributors from a variety of fields. This paper presents a tutorial on the main Gaussian filters that are used for state estimation of stochastic dynamic systems. The main concept of state estimation is firstly described based on the Bayesian paradigm and Gaussian assumption of the noise. The filters are then categorized into several groups based on their applications for state estimation. These groups involve linear optimal filtering, nonlinear filtering, adaptive filtering, and robust filtering. New advances and trends relevant to each technique are addressed and discussed in detail.
A probabilistic interpretation of set-membership filtering: Application to polynomial systems through polytopic bounding
2016, Automatica
Set-membership estimation is usually formulated in the context of set-valued calculus and no probabilistic calculations are necessary. In this paper, we show that set-membership estimation can be equivalently formulated in the probabilistic setting by employing sets of probability measures. Inference in set-membership estimation is thus carried out by computing expectations with respect to the updated set of probability measures $P$ as in the probabilistic case. In particular, it is shown that inference can be performed by solving a particular semi-infinite linear programming problem, which is a special case of the truncated moment problem in which only the zeroth order moment is known (i.e., the support). By writing the dual of the above semi-infinite linear programming problem, it is shown that, if the nonlinearities in the measurement and process equations are polynomial and if the bounding sets for initial state, process and measurement noises are described by polynomial inequalities, then an approximation of this semi-infinite linear programming problem can efficiently be obtained by using the theory of sum-of-squares polynomial optimization. We then derive a smart greedy procedure to compute a polytopic outer-approximation of the true membership-set, by computing the minimum-volume polytope that outer-bounds the set that includes all the means computed with respect to $P$ .
Robust Kalman and Bayesian Set-Valued Filtering and Model Validation for Linear Stochastic Systems
2023, SIAM-ASA Journal on Uncertainty Quantification
An Adaptive Multiple Backtracking UKF Method Based on Krein Space Theory for Marine Vehicles Alignment Process
2023, IEEE Transactions on Vehicular Technology

View all citing articles on Scopus

^☆: This work was supported by the Australian Research Council. A preliminary version of this paper was presented at the 1995 IFAC Conference on Youth Automation, Beijing. This paper was recommended for publication in revised form by Editor Peter Dorato.

View full text

Technical communiqueRobust state estimation and model validation for discrete-time uncertain systems with a deterministic description of noise and uncertainty☆

Abstract

Recursive state estimation for a set-membership description of uncertainty

IEEE Trans. Automat. Control

Minimax robust deconvolution filters under stochastic parametric and noise uncertainties

IEEE Trans. Signal Process

Technical communique
Robust state estimation and model validation for discrete-time uncertain systems with a deterministic description of noise and uncertainty☆