pith. sign in

arxiv: 2512.15419 · v2 · submitted 2025-12-17 · 💻 cs.IT · math.IT

Variational Robust Kalman Filters: A Unified Framework

Shilei Li , Dawei Shi , Hao Yu , Ling Shi This is my paper

Pith reviewed 2026-05-16 21:45 UTC · model grok-4.3

classification 💻 cs.IT math.IT
keywords variational Kalman filterrobust filteringadaptive filteringStudent's t-distributionvariational inferenceprobabilistic switchingnoise modelingstate estimation
0
0 comments X

The pith

A single variational Kalman filter unifies robustness and adaptivity by treating the former as a prerequisite for the latter via a probabilistic switching rule.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents a variational robust Kalman filter that models noise with a Student's t-distribution and uses variational inference to derive an efficient update. It shows that robustness, achieved by temporarily inflating noise covariance estimates, must precede adaptivity, which updates noise beliefs from measurements. These two goals are combined in one framework through a probabilistic switching rule that decides when to apply each behavior. By adjusting parameters the same filter can reproduce standard Kalman filtering, robust filtering, or adaptive filtering, and it handles outliers in both process and measurement noise at once. Simulations confirm better performance than separate approaches when noise is complex or time-varying.

Core claim

The central claim is that robustness can be understood as a prerequisite for adaptivity, making it possible to merge the two competing goals into a single framework through a probabilistic switching rule. The filter is built on a Student's t-distribution induced loss function solved by variational inference, and it recovers conventional, robust, and adaptive Kalman filters by parameter tuning while suppressing imperfect process and measurement noise.

What carries the argument

Variational inference on a Student's t-distribution loss function combined with a probabilistic switching rule that selects between robust inflation and adaptive update modes.

If this is right

  • The same filter recovers conventional Kalman filtering, robust Kalman filtering, and adaptive Kalman filtering simply by changing its parameters.
  • It suppresses outliers in both process noise and measurement noise within one computation.
  • Robustness acts as an enabling step that allows subsequent adaptivity to function reliably.
  • Performance improves over competing methods in environments where noise is both heavy-tailed and time-varying.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The switching rule could be tested in nonlinear state estimation problems where linear Kalman assumptions break down.
  • Similar probabilistic merging might apply to other estimation tasks that currently treat robustness and adaptation as separate design choices.
  • Real-time implementations might reduce engineering effort by eliminating the need to maintain and switch between multiple filter variants.

Load-bearing premise

The Student's t-distribution adequately models the actual noise statistics and the variational approximation together with the switching rule remains accurate and stable across the tested noise conditions.

What would settle it

An experiment in which the unified filter produces higher estimation error than separately tuned robust and adaptive filters when both process and measurement noise contain simultaneous outliers.

Figures

Figures reproduced from arXiv: 2512.15419 by Dawei Shi, Hao Yu, Ling Shi, Shilei Li.

Figure 1
Figure 1. Figure 1: The visualization of Lst and Lgau as well as their influence functions and induced PDFs. (a) The loss function of Lst and Lgau. (b) The influence function of Lst and Lgau. (c) The mapped Student’s t distribution and Gaussian distribution. (d) The PDF of latent variable λ. indicating its robustness to absolute errors that are much greater than √ ντ . According to Properties 1, 2, and 3, we have the followin… view at source ↗
Figure 2
Figure 2. Figure 2: Some noise scenarios considered in this work (but not limited to these examples). The first, second, and third column corresponds to Scenario 1, 2, and 3. The data with an absolute value bigger than 20 are visualized as ±20. (a) Case 1: wk ∼ N (0, 1), vk ∼ 0.99N (0, 1) + 0.01N (0, 400). (b) Case 2: wk ∼ N (0, 1), vk ∼ N (0, Rk,t) where Rk,t = (1+2| sin(0.1πt)|) 2 . (c) Case 3: wk ∼ N (0, 1), vk ∼ 0.99N (0,… view at source ↗
Figure 3
Figure 3. Figure 3: Error performances of VBKF-fixed, STKF, and KF. In Case 2 with adaptive measurement noise, we set the initial process and measurement covariance as Q = BBT , R = 0.1, and use ρ = 0.99 in VBKF. As in STKF-AR1, we apply the same initial process and measurement covariance as is used in VBKF. Moreover, we set ρ1 = ρ2 = 1, ρ3 = 0.99, τ 2 i = 1 for i = 1, 2, 3, ν1 = ν2 = 108 , and ν3 = 100. The estimated covaria… view at source ↗
Figure 4
Figure 4. Figure 4: Average RMSE (ARMSE) with different ν3 in STKF. 0 10 20 30 40 50 0 0.5 1 1.5 2 2.5 3 [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: The measurement noise covariance (or variance) tracking performance of VBKF and STKF-AR1. B. Example 2: Convergence Speed Investigation Following system dynamics (52), we keep ρ1 = ρ2 = 1 and investigate the effect of ρ3 = ρ by considering the following step-like measurement covariance: vk ∼    N (0, 0.1), k ≤ 2000 N (0, 2.5), 2000 < k ≤ 4000 N (0, 0.1), k ≥ 4000. (53) In the simulation, we compare th… view at source ↗
Figure 6
Figure 6. Figure 6: Theoretical (based on Theorem 5) and practical variance convergence rate and the corresponding estimation variance with different ρ. The time constant is obtained by Theorem 6, expressed in seconds. 0.9 0.92 0.94 0.96 0.98 1 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.02 0.04 0.06 0.08 0.1 0.12 (a) 0.9 0.92 0.94 0.96 0.98 1 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 (b) [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The trade-off effects of ρ in STKF-AR1. (a) The trade-off between convergence time constant and convergence variance regarding ρ. (b) The error performance with different ρ. C. Example 4: Superior Performance We consider a 1-DOF torsion load system with unknown disturbances as given in [38], [39]. The discrete system dynamics, with sampling time of dt = 0.01 and maximum time step Nt = 2000, are given by xk… view at source ↗
Figure 8
Figure 8. Figure 8: The measurement covariance (or variance) tracking performance of VBKF and STKF-AR1 in Case 2. The blue and orange lines denote the estimated variance, and the yellow line denote the ground truth variance (for both two measurement channels). (a) The performance of VBKF. (b) The performance of STKF-AR1. 0 5 10 15 20 0 2 4 6 8 10 12 (a) VBKF 0 5 10 15 20 0 0.2 0.4 0.6 0.8 1 (b) STKF-AR1 [PITH_FULL_IMAGE:figu… view at source ↗
Figure 9
Figure 9. Figure 9: The measurement covariance (or variance) tracking performance of VBKF and STKF-AR2 in Case 3. (a) The performance of VBKF. (b) The performance of STKF-AR2. V. Conclusion This work bridges the gap between the robust Kalman filter and the adaptive filter. Specifically, we prove that the STKF, derived by the Student’s t-distribution induced loss and solved by fixed-point iteration, can be understood as a prer… view at source ↗
read the original abstract

Robustness and adaptivity are two competing objectives in Kalman filters (KF). Robustness involves temporarily inflating prior estimates of noise covariances, while adaptivity updates prior beliefs by exploiting measurements. In practical applications, both process and measurement noise can be influenced by outliers, be time-varying, or both. In this work, we propose a variational robust Kalman filter, built on a Student's $t$-distribution induced loss function and variational inference, and solved in a computationally efficient manner. We demonstrate that robustness can be understood as a prerequisite for adaptivity, making it possible to merge the above two competing goals into a single framework through a probabilistic switching rule. Additionally, our proposed filter can recover conventional KF, robust KF, and adaptive KF by tuning parameters, and can suppress both the imperfect process and measurement noise, enabling it to perform superiorly in complex noise environments. Simulations verify the effectiveness of the proposed method.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper proposes a variational robust Kalman filter (VRKF) built on a Student's t-distribution induced loss function and variational inference. It claims that robustness is a prerequisite for adaptivity, allowing the two goals to be merged into a single framework via a probabilistic switching rule. The filter is asserted to recover conventional KF, robust KF, and adaptive KF by tuning parameters (including the degrees of freedom), suppress both imperfect process and measurement noise, and outperform existing methods in complex noise environments, with simulations verifying effectiveness.

Significance. If the unification via the switching rule holds with provable stability, the work would offer a principled single-framework approach to handling outliers and time-varying noise in Kalman filtering, which is valuable for applications in signal processing and control. The parameter-tuning recovery of standard methods is a positive feature that could aid adoption. However, the current lack of formal analysis on the switching mechanism and limited simulation details reduce the immediate significance.

major comments (2)
  1. [Abstract and variational inference derivation] The load-bearing claim that the probabilistic switching rule (induced by variational inference on the t-distribution loss) merges robustness and adaptivity while remaining stable lacks any derivation, bound, or analysis showing that the variational lower bound yields well-defined switching probabilities when process and measurement outliers occur concurrently and are time-varying. The construction implicitly assumes approximation error does not accumulate in the joint state-noise posterior (see abstract and the section deriving the switching rule).
  2. [Simulations section] The abstract states that simulations verify effectiveness and that the filter recovers conventional methods by tuning, but provides no derivation details, error analysis, explicit comparison baselines, or quantitative metrics (e.g., RMSE tables or stability metrics across noise conditions), leaving the central performance claims only moderately supported.
minor comments (1)
  1. [Abstract] Clarify the exact role and tuning range of the degrees of freedom parameter in the t-distribution for recovering the standard KF, robust KF, and adaptive KF cases.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and insightful comments on our manuscript. We address each major comment below and describe the revisions we will incorporate to strengthen the paper.

read point-by-point responses

  1. Referee: [Abstract and variational inference derivation] The load-bearing claim that the probabilistic switching rule (induced by variational inference on the t-distribution loss) merges robustness and adaptivity while remaining stable lacks any derivation, bound, or analysis showing that the variational lower bound yields well-defined switching probabilities when process and measurement outliers occur concurrently and are time-varying. The construction implicitly assumes approximation error does not accumulate in the joint state-noise posterior (see abstract and the section deriving the switching rule).

    Authors: We thank the referee for this observation. The probabilistic switching rule is obtained directly from the variational inference optimization of the Student's t-induced loss, where the variational posterior over the noise scaling variables yields the switching probabilities as a byproduct of the evidence lower bound. This construction is presented in the derivation section. We agree, however, that explicit bounds on the switching probabilities under concurrent time-varying outliers and a dedicated analysis of approximation error accumulation in the joint state-noise posterior are not provided. In the revision we will add a new subsection that derives such bounds from the properties of the variational approximation and discusses conditions under which the switching remains well-defined and stable. revision: yes

  2. Referee: [Simulations section] The abstract states that simulations verify effectiveness and that the filter recovers conventional methods by tuning, but provides no derivation details, error analysis, explicit comparison baselines, or quantitative metrics (e.g., RMSE tables or stability metrics across noise conditions), leaving the central performance claims only moderately supported.

    Authors: We accept that the simulation section requires expansion to more rigorously support the claims. In the revised manuscript we will: (i) provide step-by-step derivation details showing how specific parameter choices (degrees of freedom, prior noise covariances) recover the conventional KF, robust KF, and adaptive KF; (ii) include explicit comparison baselines consisting of the standard Kalman filter, representative robust KF variants, and adaptive KF methods; and (iii) report quantitative results via RMSE tables together with stability metrics (e.g., mean squared error convergence and outlier rejection rates) across a range of noise conditions, including concurrent process and measurement outliers. These additions will furnish stronger empirical evidence. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper presents a variational robust Kalman filter constructed from a Student's t-distribution loss and standard variational inference, with a probabilistic switching rule offered as the mechanism that unifies robustness and adaptivity. No load-bearing step is shown to reduce by the paper's own equations to a fitted input, self-definition, or self-citation chain; the recovery of conventional KF, robust KF, and adaptive KF is described as occurring through parameter tuning rather than by construction. The central demonstration that robustness is a prerequisite for adaptivity is framed as an interpretive consequence of the variational setup, not as a tautological renaming or imported uniqueness result. The derivation therefore remains self-contained against external benchmarks such as classical Kalman filter theory and variational inference methods.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on the Student's t-distribution for robustness, variational inference for tractable computation, and the assumption that a probabilistic switch can merge the two objectives without introducing instability.

free parameters (1)
  • degrees of freedom parameter in t-distribution
    Controls outlier robustness and is likely tuned or selected to achieve the claimed recovery of other filters.
axioms (1)
  • domain assumption Variational inference yields a sufficiently accurate approximation to the true posterior for the switching rule to function as intended
    Invoked to justify the computationally efficient solution.

pith-pipeline@v0.9.0 · 5450 in / 1200 out tokens · 51698 ms · 2026-05-16T21:45:29.096274+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

38 extracted references · 38 canonical work pages

  1. [1]

    Exactly sparse Gaussian variational inference with application to derivative-free batch nonlinear state estimation,

    T. D. Barfoot, J. R. Forbes, and D. J. Yoon, “Exactly sparse Gaussian variational inference with application to derivative-free batch nonlinear state estimation,” The International Journal of Robotics Research, vol. 39, no. 13, pp. 1473–1502, 2020

    work page 2020

  2. [2]

    Multivariate stochastic variance models,

    A. Harvey, E. Ruiz, and N. Shephard, “Multivariate stochastic variance models,” The Review of Economic Studies, vol. 61, no. 2, pp. 247–264, 1994

    work page 1994

  3. [3]

    Dynamic variational level sets for cardiac 4d reconstruction,

    A. Keil, “Dynamic variational level sets for cardiac 4d reconstruction,” Ph.D. dissertation, Technische Universität München, 2010

    work page 2010

  4. [4]

    Review of the ensemble Kalman filter for atmospheric data assimilation,

    P. L. Houtekamer and F. Zhang, “Review of the ensemble Kalman filter for atmospheric data assimilation,” Monthly Weather Review, vol. 144, no. 12, pp. 4489–4532, 2016

    work page 2016

  5. [5]

    State estimation of a physical system with unknown governing equations,

    K. Course and P. B. Nair, “State estimation of a physical system with unknown governing equations,” Nature, vol. 622, no. 7982, pp. 261–267, 2023

    work page 2023

  6. [7]

    On the a priori information in sequential estimation problems,

    T. Nishimura, “On the a priori information in sequential estimation problems,” IEEE Transactions on Automatic Control, vol. 11, no. 2, pp. 197–204, 1966

    work page 1966

  7. [8]

    H∞ design of optimal linear filters,

    M. Grimble, “ H∞ design of optimal linear filters,” Linear Circuit Systems and Signal Processing: Theory and Application (Proc. MTNS’87), pp. 533–540, 1988

    work page 1988

  8. [9]

    Huber-based novel robust unscented Kalman filter,

    L. Chang, B. Hu, G. Chang, and A. Li, “Huber-based novel robust unscented Kalman filter,” IET Science, Measurement & Technology, vol. 6, no. 6, pp. 502–509, 2012

    work page 2012

  9. [10]

    ℓ2 and ℓ1 trend filtering: A Kalman filter approach,

    A. K. Roonizi, “ ℓ2 and ℓ1 trend filtering: A Kalman filter approach,” IEEE Signal Processing Magazine, vol. 38, no. 6, pp. 137–145, 2021

    work page 2021

  10. [11]

    A novel robust nonlinear Kalman filter based on multivariate Laplace distribution,

    G. Wang, C. Yang, and X. Ma, “A novel robust nonlinear Kalman filter based on multivariate Laplace distribution,” IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 68, no. 7, pp. 2705–2709, 2021

    work page 2021

  11. [12]

    A novel robust student’s t-based Kalman filter,

    Y. Huang, Y. Zhang, N. Li, Z. Wu, and J. A. Chambers, “A novel robust student’s t-based Kalman filter,” IEEE Transactions on Aerospace and Electronic Systems, vol. 53, no. 3, pp. 1545–1554, 2017

    work page 2017

  12. [13]

    J. C. Principe, Information theoretic learning: Renyi’s entropy and kernel perspectives. Springer Science & Business Media, 2010

    work page 2010

  13. [14]

    Vapnik, The nature of statistical learning theory

    V. Vapnik, The nature of statistical learning theory. Springer science & business media, 2013

    work page 2013

  14. [15]

    Multi-kernel maximum correntropy Kalman filter,

    S. Li, D. Shi, W. Zou, and L. Shi, “Multi-kernel maximum correntropy Kalman filter,” IEEE Control Systems Letters, vol. 6, pp. 1490–1495, 2021

    work page 2021

  15. [16]

    Maximum correntropy Kalman filter,

    B. Chen, X. Liu, H. Zhao, and J. C. Principe, “Maximum correntropy Kalman filter,” Automatica, vol. 76, pp. 70–77, 2017

    work page 2017

  16. [17]

    Generalized multi-kernel maximum correntropy Kalman filter for disturbance estimation,

    S. Li, D. Shi, Y. Lou, W. Zou, and L. Shi, “Generalized multi-kernel maximum correntropy Kalman filter for disturbance estimation,” IEEE Transactions on Automatic Control, 2023

    work page 2023

  17. [18]

    Statistical similarity measure-based adaptive outlier-robust state estimator with applications,

    M. Bai, Y. Huang, Y. Zhang, and J. Chambers, “Statistical similarity measure-based adaptive outlier-robust state estimator with applications,” IEEE Transactions on Automatic Control, vol. 67, no. 8, pp. 4354–4361, 2022

    work page 2022

  18. [19]

    Chandrasekhar-based maximum correntropy Kalman filtering with the adaptive kernel size selection,

    M. V. Kulikova, “Chandrasekhar-based maximum correntropy Kalman filtering with the adaptive kernel size selection,” IEEE Transactions on Automatic Control, vol. 65, no. 2, pp. 741–748, 2019

    work page 2019

  19. [20]

    Using correntropy as a cost function in linear adaptive filters,

    A. Singh and J. C. Principe, “Using correntropy as a cost function in linear adaptive filters,” in 2009 International Joint Conference on Neural Networks. IEEE, 2009, pp. 2950–2955

    work page 2009

  20. [21]

    Convergence of a fixed-point algorithm under maximum correntropy criterion,

    B. Chen, J. Wang, H. Zhao, N. Zheng, and J. C. Principe, “Convergence of a fixed-point algorithm under maximum correntropy criterion,” IEEE Signal Processing Letters, vol. 22, no. 10, pp. 1723–1727, 2015

    work page 2015

  21. [22]

    Maximum likelihood identification of stochastic linear systems,

    R. Kashyap, “Maximum likelihood identification of stochastic linear systems,” IEEE Transactions on Automatic Control, vol. 15, no. 1, pp. 25–34, 1970

    work page 1970

  22. [23]

    Covariance matching based adaptive unscented Kalman filter for direct filtering in ins/gnss integration,

    Y. Meng, S. Gao, Y. Zhong, G. Hu, and A. Subic, “Covariance matching based adaptive unscented Kalman filter for direct filtering in ins/gnss integration,” Acta Astronautica, vol. 120, pp. 171–181, 2016

    work page 2016

  23. [24]

    Recursive noise adaptive Kalman filtering by variational bayesian approximations,

    S. Sarkka and A. Nummenmaa, “Recursive noise adaptive Kalman filtering by variational bayesian approximations,” IEEE Transactions on Automatic control, vol. 54, no. 3, pp. 596–600, 2009

    work page 2009

  24. [25]

    A novel adaptive Kalman filter with inaccurate process and measurement noise covariance matrices,

    Y. Huang, Y. Zhang, Z. Wu, N. Li, and J. Chambers, “A novel adaptive Kalman filter with inaccurate process and measurement noise covariance matrices,” IEEE transactions on Automatic Control, vol. 63, no. 2, pp. 594–601, 2017

    work page 2017

  25. [26]

    Variational bayesian unscented Kalman filter for active distribution system state estimation,

    D. Ćetenović, J. Zhao, V. Levi, Y. Liu, and V. Terzija, “Variational bayesian unscented Kalman filter for active distribution system state estimation,” IEEE Transactions on Power Systems, 2024

    work page 2024

  26. [27]

    Variational bayesian adaptive cubature information filter based on wishart distribution,

    P. Dong, Z. Jing, H. Leung, and K. Shen, “Variational bayesian adaptive cubature information filter based on wishart distribution,” IEEE Transactions on Automatic Control, vol. 62, no. 11, pp. 6051–6057, 2017

    work page 2017

  27. [28]

    Generalized Kalman smoothing: Modeling and algorithms,

    A. Aravkin, J. V. Burke, L. Ljung, A. Lozano, and G. Pillonetto, “Generalized Kalman smoothing: Modeling and algorithms,” Automatica, vol. 86, pp. 63–86, 2017

    work page 2017

  28. [29]

    Generalized correntropy for robust adaptive filtering,

    B. Chen, L. Xing, H. Zhao, N. Zheng, J. C. Prı et al., “Generalized correntropy for robust adaptive filtering,” IEEE Transactions on Signal Processing, vol. 64, no. 13, pp. 3376–3387, 2016

    work page 2016

  29. [30]

    Variational bayesian adaptation of process noise covariance matrix in Kalman filtering,

    G. Chang, C. Chen, Q. Zhang, and S. Zhang, “Variational bayesian adaptation of process noise covariance matrix in Kalman filtering,” Journal of the Franklin Institute, vol. 358, no. 7, pp. 3980–3993, 2021

    work page 2021

  30. [31]

    Kalman filter with both adaptivity and robustness,

    G. Chang, “Kalman filter with both adaptivity and robustness,” Journal of Process Control, vol. 24, no. 3, pp. 81–87, 2014

    work page 2014

  31. [32]

    A variational bayesian-based unscented Kalman filter with both adaptivity and robustness,

    K. Li, L. Chang, and B. Hu, “A variational bayesian-based unscented Kalman filter with both adaptivity and robustness,” IEEE Sensors Journal, vol. 16, no. 18, pp. 6966–6976, 2016

    work page 2016

  32. [33]

    Gelman, J

    A. Gelman, J. B. Carlin, H. S. Stern, and D. B. Rubin, Bayesian data analysis. Chapman and Hall/CRC, 1995

    work page 1995

  33. [34]

    Outlier models and prior distributions in bayesian linear regression,

    M. West, “Outlier models and prior distributions in bayesian linear regression,” Journal of the Royal Statistical Society Series B: Statistical Methodology, vol. 46, no. 3, pp. 431–439, 1984

    work page 1984

  34. [35]

    Gaussian process regression with student’s t likelihood,

    J. Vanhatalo, P. Jylänki, and A. Vehtari, “Gaussian process regression with student’s t likelihood,” Advances in Neural Information Processing Systems, vol. 22, 2009

    work page 2009

  35. [36]

    T. D. Barfoot, State estimation for robotics. Cambridge University Press, 2024

    work page 2024

  36. [37]

    B. G. Liptak, Instrument Engineers’ Handbook, Volume One: Process Measurement and Analysis. CRC press, 2003

    work page 2003

  37. [38]

    Fusion Kalman/UFIR filter for state estimation with uncertain parameters and noise statistics,

    S. Zhao, Y. S. Shmaliy, P. Shi, and C. K. Ahn, “Fusion Kalman/UFIR filter for state estimation with uncertain parameters and noise statistics,” IEEE Transactions on Industrial Electronics, vol. 64, no. 4, pp. 3075–3083, 2016

    work page 2016

  38. [39]

    A Kalman & fading memory co-filter for uncertain systems based on self-perception mechanism,

    X. Luan, W. Xue, S. Zhao, and F. Liu, “A Kalman & fading memory co-filter for uncertain systems based on self-perception mechanism,” IEEE Transactions on Automatic Control, 2025. 23 PLACE PHOTO HERE Shilei Li received the B.E. degree in Detection Guidance and Control Technology and M.S. degree in Control Engineering both from Harbin Institute of Technolog...

    work page 2025