Pith. sign in

REVIEW 4 major objections 6 minor 74 references

An adaptive split-combine Gaussian mixture filter for nonlinear and multimodal state estimation

T0 review · 4 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A new Gaussian mixture filter, AMF, adaptively splits and merges Gaussian particles during state propagation, and its splitting rule provably halves variance along the exact direction where the local linear approximation fails, letting it…

desk verdict A promising adaptive split-combine GMF with a clean variance-reduction theorem, but the algorithm as written has a square-root inconsistency that needs fixing before it is executable. read the letter →

arxiv 2608.04430 v3 pith:5CKU4PGD submitted 2026-08-05 math.NA cs.NAstat.ML

San Kim , Won Chang , Daniel B. Forger , Dae Wook Kim This is my paper · ORCID
classification math.NAcs.NAstat.ML MSC 60G3562M2065C2093E11
keywords Gaussianmixturefilteringadaptivesplittinglevel-setpropagationnonlinearstateestimationmultimodaldensitiesVanderPoloscillatorLorenzattractorparallelcomputation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper develops an adaptive split-combine Gaussian mixture filter (AMF) for Bayesian state estimation in nonlinear and multimodal systems. The filter propagates each Gaussian particle by tracking its level set under a local linear approximation, splits particles when that approximation breaks, and merges particles when a KL-divergence upper bound says the merge is safe. Its central claim is a splitting rule that provably halves the variance along the target level-set-point direction, the direction where the linearization error is worst, rather than along an eigenvector as in prior work. The authors argue this keeps particle counts low, avoids online numerical optimization, and makes the filter accurate on Van der Pol oscillators, coupled oscillators, and the Lorenz attractor while remaining parallelizable.

What carries the argument

The load-bearing machinery is the variance-halving split formula $\Sigma_{\mathrm{split}} = \Sigma - \tfrac{1}{2}\Sigma_{\parallel}$, with $\Sigma_{\parallel} = (\Sigma d)(\Sigma d)^T/(d^T\Sigma d)$, obtained via a linear transformation $T = [d^T; B]^T$ that makes the target direction a principal axis. Supporting it are the local-linearity error (13) that decides when a split is needed, the level-set propagation ODEs (11)-(12) that move each Gaussian particle, and the KL-divergence upper bound (39) that decides when two particles can be safely combined.

What would settle it

Take the Van der Pol time-update experiment (µ=2.0) from the paper and instrument the code: after each particle split, recompute the covariance square root M (e.g., by Cholesky) and evaluate the linearity error (13) along a direction obtained by rotating one of the column vectors, not just along the original columns. If that error does not strictly decrease when a split was triggered, or if the propagated PDF deviates from the 50,000-particle particle filter ground truth beyond the error bars reported in Fig. 5, then the variance-halving guarantee does not translate into the accuracy improvement the paper claims.

Watch

Extended reading notes

Core claim

The paper's central discovery is a geometric splitting identity: for a Gaussian particle with covariance $\Sigma$ and a target direction $d$ pointing to a level-set point, one can construct a linear transformation $T$ that decorrelates $d$ from the other coordinates, split the transformed Gaussian along that principal axis using the precomputed one-dimensional optimal parameters $a = 1.03332\sigma$ and $w = 0.21921$, and transform back. The resulting child covariance is $\Sigma_{\mathrm{split}} = \Sigma - \tfrac{1}{2}\Sigma_{\parallel}$ with $\Sigma_{\parallel} = (\Sigma d)(\Sigma d)^T/(d^T\Sigma d)$, so the variance along $d$ is exactly halved (Theorem 2.1). Because the split is driven by the same local-linearity error detector (13) used during propagation, it attacks the actual failure mode, unlike eigenvector-based splitting, and the KL-upper-bound combining rule (39) prevents harmful merges. The paper reports that AMF consistently outperforms CD-CKF, LSKF, GS-ACD-ECKF, and APPDF on the benchmarks, using fewer particles than APPDF and running faster in parallel.

Load-bearing premise

The filter's accuracy rests on the assumption that checking the linearization error along just the d column directions of each particle's covariance is enough to find where the Gaussian approximation breaks, and that the particle's covariance square root can always be recovered after a split; if those fail, the level-set propagation that carries the method silently stops being Gaussian.

Editorial extensions

If this is right

  • If AMF's central claim holds, non-Gaussian filtering can be performed accurately with a small, adaptively sized Gaussian mixture, because each split directly reduces the linearization error that would otherwise corrupt level-set propagation.
  • The split formula needs only the level-set direction and the covariance square root, so it avoids online optimization and matrix inversions, making the filter cheap enough to run in parallel across particles.
  • The paper's diagnosis of APPD's failure mode—eigenvector splitting causing excessive splits, pruning, and information loss—implies that direction-aware splitting should outperform eigenvector-based splitting in any Gaussian-mixture uncertainty propagation setting.
  • The method's accuracy on nonlinear oscillators and the Lorenz attractor suggests it could replace particle filters in data-assimilation problems where the state dimension is moderate and strong nonlinearity makes single-Gaussian filters unreliable.
  • The cubature-based measurement update and parallel particle updates imply AMF can scale to higher-dimensional coupled oscillator networks, with the main bottleneck being the Gaussian assumption of the per-particle measurement update.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The variance-halving identity is purely algebraic and does not depend on the filtering context; the same split rule could be transplanted into any Gaussian-mixture uncertainty propagation or generative modeling pipeline that tracks probability densities over time.
  • The paper checks the local-linearity error (13) only along the d column directions of M; a natural test is whether evaluating it on a denser set of directions, or on the principal eigenvector, changes the split decisions and further reduces particle counts without hurting accuracy.
  • Algorithm 2 outputs a covariance matrix, while Algorithm 1 requires its square root M; the paper never states how M is recovered after a split. A practical implementation would need Cholesky or a matrix-square-root update, and if that recovery is unstable, the theoretical variance guarantee may not be realized numerically.
  • The KL-upper-bound combining is applied cell-by-cell on a spatial grid; choosing the grid adaptively or using a global greedy combining might produce smaller mixtures or better preservation of multimodality, though at higher computational cost.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. This paper develops an adaptive split-combine Gaussian mixture filter (AMF) for continuous-discrete nonlinear filtering. The time update propagates each Gaussian particle using the level-set Kalman filter (LSKF), splits particles when a relative-error criterion indicates breakdown of the local linear approximation, and combines them using a computable upper bound on the KL divergence. The measurement update applies a cubature-rule correction to each particle and updates weights by numerical marginal likelihoods. The central theoretical contribution is Theorem 2.1, which shows that the proposed splitting reduces the covariance along any chosen level-set-point direction d by subtracting half of (Sigma d)(Sigma d)^T / (d^T Sigma d). Numerical experiments on single and coupled Van der Pol oscillators and the Lorenz attractor compare AMF against CD-CKF, LSKF, GS-ACD-ECKF, and APPDF, reporting improved accuracy and reduced particle counts.

Significance. If the construction in Theorem 2.1 and the KL upper bound are used as stated, the method is a useful and largely optimization-free alternative to existing adaptive Gaussian mixture filters. The variance-reduction identity is proved cleanly, the KL combining criterion is a bound rather than an ad hoc proximity rule, and the benchmark suite covers genuinely nonlinear and chaotic regimes. The paper would be considerably strengthened by released code and a fully specified numerical procedure, but the central splitting identity itself is a solid contribution.

major comments (4)
  1. [Sec. 2.3.2, Eq. (14)] The covariance produced by the explicit square-root formula in Eq. (14) is not the covariance of Theorem 2.1 in the non-eigenvector case. Writing c = 1 - 1/sqrt(2) and u = x_i / ||x_i||, Eq. (14) defines N = (I - c u u^T)M, so NN^T = Sigma - c(Sigma u u^T + u u^T Sigma) + c^2 (u^T Sigma u) u u^T. This equals Sigma - (1/2)(Sigma d)(Sigma d)^T / (d^T Sigma d) only when u is an eigenvector of Sigma. Since the paper's motivation is precisely that the level-set direction is generally not aligned with a principal axis, Eq. (14) cannot be the square-root construction for the split used in Algorithm 2. The manuscript should either delete Eq. (14) or clearly restrict it to the eigenvector case; otherwise a reader implementing the stated formula will propagate a covariance different from the one covered by Theorem 2.1.
  2. [Algorithm 1, Algorithm 2] Algorithm 1 requires a covariance square root M, but Algorithm 2 returns only Sigma_split (line 13) and gives no method to obtain a factor M_child with M_child M_child^T = Sigma_split. The ODE in Eq. (11) cannot be integrated from Sigma alone. A Cholesky factorization of Sigma_split would close the loop, but it is not specified, and it is not covered by the claimed avoidance of explicit matrix operations. This is a blocking implementation gap that prevents a reader from executing the filter as written.
  3. [Sec. 2.3.2, Eq. (13), Algorithm 2 lines 4-7] The linearity check evaluates epsilon only at Delta x = M_i for i = 1,...,d. Nothing ensures that the worst nonlinearity over the support of the Gaussian is sampled by these d columns; after a split or a propagation step the columns are no longer tied to the level-set directions along which nonlinearity may have developed. The claim that the method 'guarantees' satisfaction of the local linear approximation is therefore stronger than what the criterion can certify. A concrete test would be to compare the decision from Eq. (13) at the columns with the maximum of Eq. (13) over random unit directions scaled by the local standard deviation.
  4. [Sec. 3.2 (Example 2), p. 30] The measurement values in Example 2 are explicitly selected to deviate from the ground-truth limit-cycle trajectory in a particular way. This single hand-picked realization makes it difficult to assess whether AMF's advantage over the baselines is robust to measurement noise. Reporting results for one fixed measurement path is not sufficient for the claim that AMF 'consistently outperforms' across the benchmark; at minimum a sensitivity analysis over multiple measurement noise realizations is needed.
minor comments (6)
  1. [Sec. 2.3.2, Eq. (13)] The denominator ||v(xbar)|| in Eq. (13) can vanish near fixed points or slow manifolds, making the relative error undefined and the splitting decision ill-posed in exactly the regimes the method targets.
  2. [Fig. 5(g), Fig. 6(g), Sec. 3.1] The main text sets the process noise as K = 0.006 I_2, while the captions of Figs. 5(g) and 6(g) report K = 0.06 I_2 for the same experiment. One of these is a typo and should be corrected.
  3. [Sec. 3.2, Eq. (50)] The measurement noise matrix is displayed as R with entries "0.52" and "0.8 2"; presumably these are meant to be 0.5^2 and 0.8^2. The notation should be cleaned up.
  4. [Algorithm 4] The symbol N is reused for the concatenated cubature-point matrix in Algorithm 4, whereas N denotes the split covariance square root in Eq. (14). This overloading is confusing and should be avoided.
  5. [Sec. 2.3.3, Eq. (37)] In Eq. (37) the symbol pi_com is used without stating whether it is the unweighted Gaussian density N(mu_com, Sigma_com) or the weighted density w_com N(mu_com, Sigma_com). The derivation in Appendix D assumes the latter, but the main text reads as the former; the notation should be made consistent.
  6. [Sec. 6] The data availability statement says code will be released only upon acceptance. For a methods paper whose central claim depends on a specific algorithmic flow, making the code available at review time or providing a complete, self-contained pseudocode for the square-root recovery would substantially improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the variance-reduction theorem and KL bound are proved from explicit constructions; benchmarks use external PF ground truth.

full rationale

The central claim, Theorem 2.1, is a proved algebraic identity: starting from an explicitly constructed linear transformation T that makes the target direction d uncorrelated with the remaining coordinates, the child covariance is computed as Sigma - (1/2)Sigma_parallel, where Sigma_parallel = (Sigma d)(Sigma d)^T / (d^T Sigma d). This is a direct calculation (Eqs. 17-34), not a fitted parameter or renamed input. The 1-D split parameters a=1.03332 sigma and w=0.21921 are obtained from the fixed minimax problem (16), and Appendix C proves the multivariate splitting error equals the 1-D error via independence of y1 and y_perp; they are not fitted to the benchmark PDFs. The KL combining bound (39) is derived in Appendix D from standard lemmas. Self-citations to [42] and [47] provide threshold, split parameters, and single-particle propagation, but the variance-reduction guarantee is self-contained and does not depend on them. The paper's own limitations are honest scoping statements, not circularity. The inconsistency between Eq. (14) and Theorem 2.1/Algorithm 2 is a correctness/reproducibility issue, not circularity, because no quantity is defined in terms of the result it purports to predict. Therefore score is 0.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The central theorem is parameter-free, but the working filter depends on heuristic thresholds (epsilon, W, C, grid volume) and on the LSKF local-linearity assumption. The 1D split constants a and w are derived constants from a fixed minimax problem, not fitted to benchmark output. No new physical entities are introduced.

free parameters (5)
  • epsilon_threshold = 0.05
    Splitting threshold in (13), adopted from Wang and Forger (Ref. [42]) without re-derivation for the new directional split.
  • W_threshold = 1e-4 (1e-5 for Lorenz)
    Pruning threshold; chosen so a particle with weight 1 can be split at most 20 times.
  • C_threshold = 0.03 (Van der Pol), 0.05 (coupled Van der Pol, Lorenz)
    KL-bound combining threshold; selected per benchmark based on effective support size.
  • grid_cell_volume = 0.04^2 (VdP), 1^6 (coupled), 1^3 (Lorenz)
    State-space gridding for combining; chosen per benchmark.
  • combining_radius_r (APPDF baseline) = 0.005 (VdP), 1 (coupled), 1 (Lorenz)
    Baseline hyperparameter; affects the fairness of APPDF comparison.
assumptions (4)
  • domain assumption Local linear approximation v(x) approximately equal to J x of the drift is valid for propagating each Gaussian particle's level set (Eq. 9).
    Basis of the level-set filter from Ref. [47]; the paper inherits it and relies on it for Gaussianity of particles.
  • ad hoc to paper The relative error epsilon in (13), with fixed threshold 0.05, is a sufficient indicator of when the local linear approximation breaks down.
    No theorem links epsilon to the actual propagation error; the threshold is ported from APPD.
  • standard math For Gaussian vectors, uncorrelatedness implies independence, used to factor the splitting error in Appendix C.
    Classical property; applied to justify 1D parameters.
  • standard math The KL divergence lemmas in Appendix D (joint convexity) hold for the unnormalized mixture expressions.
    Standard KL properties cited from Refs. [46,48].

how reviews work

0 comments
Cite this review

Pith. "Pith review of An adaptive split-combine Gaussian mixture filter for nonlinear and multimodal state estimation." pith.science (2026). https://pith.science/paper/5CKU4PGD

@misc{pith2026260804430,
  author       = {Pith},
  title        = {Pith review of: An adaptive split-combine Gaussian mixture filter for nonlinear and multimodal state estimation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5CKU4PGD}},
  note         = {Machine review of arXiv:2608.04430}
}
read the original abstract

Filtering combines model predictions with measurements to estimate the probability density function (PDF) of a system state over time. The PDF often becomes highly asymmetric and even multimodal in nonlinear systems with oscillatory or chaotic dynamics. Such non-Gaussian features violate the single-Gaussian assumption underlying Kalman-type filters. To address this problem, Gaussian mixture filtering has been proposed. However, accurately propagating mixture components and adaptively adjusting their number and weights over time remain open challenges. Here, we develop an adaptive split-combine Gaussian mixture filter (AMF) that estimates the time evolution of asymmetric and multimodal PDFs by adaptively splitting and combining Gaussian particles without auxiliary online numerical optimization. Notably, the proposed splitting method guarantees a reduction in variance along a target level-set-point direction of a Gaussian particle. This enables accurate and efficient propagation of particles. We show that AMF consistently outperforms various baseline filters across diverse benchmarks, including single and coupled slow-fast Van der Pol oscillators and the Lorenz attractor. We also propose a parallel implementation of AMF, allowing high-fidelity PDF estimation with practical computational cost.

Figures

Figures reproduced from arXiv: 2608.04430 by the authors.

Figure 1
Figure 1. Overview of the adaptive split–combine Gaussian mixture filter. In the time-update step, single-particle propagation is performed by tracking its level set. Particles violating the local linear approximation are split at each infinitesimal time step without auxiliary online numerical optimization, guaranteeing halved variance along the target level-set-point direction. Particles are combined when the combined distri… view at source ↗
Figure 2
Figure 2. Schematic diagram of the particle splitting method. To reduce the variance along a specific direction d, a linear transformation T is first applied to the particle so that the variance of the transformed particle along its principal axis equals the variance of the original particle along d, i.e., σ 2 d = d T Σd. The particle is then split into three child particles along the principal axis using the parameter values… view at source ↗
Figure 3
Figure 3. Particle-combining decisions based solely on particle means can be inaccurate, leading to numerical errors and unnecessary computational cost. (a) When the means of particles 1 (red) and 2 (blue) are close but their variances differ substantially, the PDF of the combined particle (purple) can differ significantly from the weighted mixture PDF (green). Thus, combining particles solely because their means are close ca… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Particle combining based on the Kullback–Leibler divergence upper bound. The upper bound in (39) is evaluated for each particle pair. If the minimum value among all pairs is below a threshold, the corresponding pair is combined, as illustrated for particles 1 and 2. 2.…
Figure 5
Figure 5. Figure 5: Performance of AMF for uncertainty propagation in the Van der Pol oscillator. (a) Mean state trajectories of AMF, the baselines, and the PF ground truth (GT) for µ = 1.6 over t = 20–27. (b) Mean trajectories of states y and z over time for µ = 1.6. (c) SSE between the …
Figure 6
Figure 6. Figure 6: Performance of AMF for filtering in the Van der Pol oscillator. (a) Mean state trajectories of AMF, the baselines, and the PF ground truth (GT) for µ = 1.5 over t = 20–27. (b) Mean trajectories of states y and z over time for µ = 1.5. (c) SSE between the estimated mean…
Figure 7
Figure 7. Figure 7: Accuracy of AMF for filtering in the coupled Van der Pol oscillators (a) Mean state trajectories of AMF, the baselines, and the PF ground truth (GT) for µ = 2.0 and k = 0.3 over t = 0.5–10.0. (b) SSE between the estimated mean trajectories and GT across different combi…
Figure 8
Figure 8. Figure 8: Changes in the number of particles due to particle splitting, prun￾ing, and combining. (a) Schematic diagram illustrating the evolution of the particle number during uncertainty propagation using AMF and APPDF. Changes in the particle number due to splitting, pruning, …
Figure 9
Figure 9. Figure 9: Performance of AMF for filtering in the Lorenz attractor. (a)–(d) Heat maps of the state PDFs projected onto the xy-, yz-, and zx-planes for PF (a), LSKF (b), APPDF (c), and AMF (d), respectively, at t = 12. (e) Marginal PDFs of states x, y, and z at t = 12. (f) SSE be…

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 9 canonical work pages

  1. [1]

    E. N. Lorenz, Deterministic nonperiodic flow 1, in: Universality in Chaos, 2nd edition, Routledge, 2017, pp. 367–378

    2017

  2. [2]

    A. L. Hodgkin, A. F. Huxley, B. Katz, Measurement of current-voltage relations in the membrane of the giant axon of loligo, The Journal of physiology 116 (4) (1952) 424

    1952

  3. [3]

    Van Der Pol, J

    B. Van Der Pol, J. Van Der Mark, Lxxii. the heartbeat considered as a relaxation oscillation, and an electrical model of the heart, The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science 6 (38) (1928) 763–775

    1928

  4. [4]

    M. E. Jewett, R. E. Kronauer, Refinement of limit cycle oscillator model of the effects of light on the human circadian pacemaker, Journal of Theoretical Biology 192 (4) (1998) 455–465

    1998

  5. [5]

    J. E. Stone, S. Postnova, T. L. Sletten, S. M. Rajaratnam, A. J. Phillips, Computational approaches for individual circadian phase prediction in field settings, Current Opinion in Systems Biology 22 (2020) 39–51

    2020

  6. [6]

    D.-J. Dijk, J. F. Duffy, Novel approaches for assessing circadian rhyth- micity in humans: a review, Journal of biological rhythms 35 (5) (2020) 421–438

    2020

  7. [7]

    D. W. Kim, C. Chang, X. Chen, A. C. Doran, F. Gaudreault, T. Wager, G. J. DeMarco, J. K. Kim, Systems approach reveals photosensitivity and per 2 level as determinants of clock-modulator efficacy, Molecular systems biology 15 (7) (2019) e8838. 39

    work page 2019

  8. [8]

    Särkkä, L

    S. Särkkä, L. Svensson, Bayesian filtering and smoothing, Vol. 17, Cam- bridge university press, 2023

    2023

Show all 74 references
  1. [9]

    Arasaratnam, S

    I. Arasaratnam, S. Haykin, Cubature kalman filters, IEEE Transactions on automatic control 54 (6) (2009) 1254–1269

    2009

  2. [10]

    Arasaratnam, S

    I. Arasaratnam, S. Haykin, T. R. Hurd, Cubature kalman filtering for continuous-discretesystems: theoryandsimulations, IEEETransactions on Signal Processing 58 (10) (2010) 4977–4993

    2010

  3. [11]

    N. J. Gordon, D. J. Salmond, A. F. Smith, Novel approach to nonlinear/non-gaussian bayesian state estimation, in: IEE proceedings F (radar and signal processing), Vol. 140, IET, 1993, pp. 107–113

    1993

  4. [12]

    Doucet, N

    A. Doucet, N. De Freitas, N. Gordon, An introduction to sequential monte carlo methods, in: Sequential Monte Carlo methods in practice, Springer, 2001, pp. 3–14

    2001

  5. [13]

    Ristic, S

    B. Ristic, S. Arulampalam, N. Gordon, Beyond the Kalman filter: Par- ticle filters for tracking applications, Artech house, 2003

    2003

  6. [14]

    Fearnhead, H

    P. Fearnhead, H. R. Künsch, Particle filters and data assimilation, An- nual Review of Statistics and Its Application 5 (2018) 421–449

    2018

  7. [15]

    Khodarahmi, V

    M. Khodarahmi, V. Maihami, A review on kalman filter models, Archives of Computational Methods in Engineering 30 (1) (2023) 727– 747

    2023

  8. [16]

    F. Daum, J. Huang, Curse of dimensionality and particle filters, in: 2003 IEEE aerospace conference proceedings (Cat. No. 03TH8652), Vol. 4, IEEE, 2003, pp. 4_1979–4_1993

    2003

  9. [17]

    Snyder, T

    C. Snyder, T. Bengtsson, P. Bickel, J. Anderson, Obstacles to high- dimensional particle filtering, Monthly Weather Review 136 (12) (2008) 4629–4640

    2008

  10. [18]

    Terejanu, P

    G. Terejanu, P. Singla, T. Singh, P. D. Scott, A novel gaussian sum filter method for accurate solution to the nonlinear filtering problem, in: 2008 11th International Conference on Information Fusion, IEEE, 2008, pp. 1–8. 40

    2008

  11. [19]

    D. W. Kim, M. P. Lee, D. B. Forger, Wearable data assimilation to estimate the circadian phase, SIAM Journal on Applied Mathematics 84 (3) (2023) S452–S475

    2023

  12. [20]

    H. W. Sorenson, D. L. Alspach, Recursive bayesian estimation using gaussian sums, Automatica 7 (4) (1971) 465–479

    1971

  13. [21]

    W. Shao, Z. Ge, Z. Song, Semisupervised bayesian gaussian mixture models for non-gaussian soft sensor, IEEE Transactions on Cybernetics 51 (7) (2019) 3455–3468

    2019

  14. [22]

    Zhang, Y

    B. Zhang, Y. C. Shin, A gaussian mixture filter with adaptive refinement for nonlinear state estimation, Signal Processing 201 (2022) 108677

    2022

  15. [23]

    P. H. Leong, S. Arulampalam, T. A. Lamahewa, T. D. Abhayapala, A gaussian-sum based cubature kalman filter for bearings-only tracking, IEEE Transactions on Aerospace and Electronic Systems 49 (2) (2013) 1161–1176

    2013

  16. [24]

    Tuggle, R

    K. Tuggle, R. Zanetti, Automated splitting gaussian mixture nonlin- ear measurement update, Journal of Guidance, Control, and Dynamics 41 (3) (2018) 725–734

    2018

  17. [25]

    Raitoharju, Á

    M. Raitoharju, Á. F. García-Fernández, S. Särkkä, Partitioned update binomial gaussian mixture filter, in: 2019 22th International Conference on Information Fusion (FUSION), IEEE, 2019, pp. 1–8

    2019

  18. [26]

    Alspach, H

    D. Alspach, H. Sorenson, Nonlinear bayesian estimation using gaussian sum approximations, IEEE transactions on automatic control 17 (4) (2003) 439–448

    2003

  19. [27]

    K. Ito, K. Xiong, Gaussian filters for nonlinear filtering problems, IEEE transactions on automatic control 45 (5) (2002) 910–927

    2002

  20. [28]

    R. Chen, J. S. Liu, Mixture kalman filters, Journal of the Royal Statis- tical Society: Series B (Statistical Methodology) 62 (3) (2000) 493–508

    2000

  21. [29]

    Faubel, J

    F. Faubel, J. McDonough, D. Klakow, The split and merge unscented gaussian mixture filter, IEEE Signal Processing Letters 16 (9) (2009) 786–789. 41

    2009

  22. [30]

    Faubel, D

    F. Faubel, D. Klakow, Further improvement of the adaptive level of detail transform: Splitting in direction of the nonlinearity, in: 2010 18th European Signal Processing Conference, 2010, pp. 850–854

    2010

  23. [31]

    Leutnant, A

    V. Leutnant, A. Krueger, R. Haeb-Umbach, A versatile Gaussian split- ting approach to non-linear state estimation and its application to noise- robust ASR, in: Interspeech 2011, 2011, pp. 1641–1644

    2011

  24. [32]

    Terejanu, P

    G. Terejanu, P. Singla, T. Singh, P. D. Scott, Uncertainty propagation for nonlinear dynamic systems using gaussian mixture models, Journal of guidance, control, and dynamics 31 (6) (2008) 1623–1633

    2008

  25. [33]

    Singla, T

    P. Singla, T. Singh, A gaussian function network for uncertainty prop- agation through nonlinear dynamical system, in: 18th Annual Space Flight Mechanics Meeting, 2008, pp. 851–864

    2008

  26. [34]

    J. T. Horwood, A. B. Poore, Adaptive gaussian sum filters for space surveillance, IEEE Transactions on Automatic Control 56 (8) (2011) 1777–1790

    2011

  27. [35]

    H. Wang, Y. Wang, C. Li, J. Li, Q. Li, X. Ban, Adaptive weight update algorithm for target tracking of uuv based on improved gaussian mixture cubature kalman filter, Complexity 2020 (2020) 1–12

    2020

  28. [36]

    Cun, An efficient adaptive weights update scheme for a gaussian mixture filter, in: 2021 33rd Chinese Control and Decision Conference (CCDC), 2021, pp

    L. Cun, An efficient adaptive weights update scheme for a gaussian mixture filter, in: 2021 33rd Chinese Control and Decision Conference (CCDC), 2021, pp. 7337–7342

    2021

  29. [37]

    Vittaldev, R

    V. Vittaldev, R. P. Russell, R. Linares, Spacecraft uncertainty propaga- tion using gaussian mixture models and polynomial chaos expansions, Journal of Guidance, Control, and Dynamics 39 (12) (2016) 2615–2626

    2016

  30. [38]

    B. A. Jones, Physics-informed domain splitting for orbit uncertainty propagation, in: AIAA SCITECH 2024 Forum, AIAA SciTech Forum, American Institute of Aeronautics and Astronautics, 2024

    2024

  31. [39]

    Duník, O

    J. Duník, O. Straka, B. Noack, J. Steinbring, U. D. Hanebeck, Direc- tional splitting of gaussian density in non-linear random variable trans- formation, IET Signal Processing 12 (9) (2018) 1073–1081. 42

    2018

  32. [40]

    Gutierrez, K

    J. Gutierrez, K. Hill, E. L. Jenson, D. J. Scheeres, J. C. Bruer, R. D. Coder, Classifying State Uncertainty for Earth-Moon Trajectories, Jour- nal of the Astronautical Sciences 71 (3) (2024) 29

    2024

  33. [41]

    Kulik, K

    J. Kulik, K. A. LeGrand, Nonlinearity- and uncertainty-informed moment-matching gaussian mixture splitting, IEEE Transactions on Aerospace and Electronic Systems 62 (2026) 1774–1794

    2026

  34. [42]

    N. Wang, D. B. Forger, The asymmetric particle population density method for simulation of coupled noisy oscillators, Journal of Compu- tational Physics 488 (2023) 112157

    2023

  35. [43]

    K. J. DeMars, R. H. Bishop, M. K. Jah, Entropy-based approach for un- certainty propagation of nonlinear dynamical systems, Journal of Guid- ance, Control, and Dynamics 36 (4) (2013) 1047–1057

    2013

  36. [44]

    Y. Liu, K. Dong, H. Wang, J. Liu, Y. He, L. Pan, Adaptive gaussian sum squared-root cubature kalman filter with split-merge scheme for state estimation, Chinese Journal of Aeronautics 27 (5) (2014) 1242– 1250

    2014

  37. [45]

    A. B. Berchet, A. Beaudoin, S. H. Huberson, Adaptive particle method based on moments for simulating the mass transport in natural flows, Computational Particle Mechanics 8 (3) (2021) 525–534

    2021

  38. [46]

    A. R. Runnalls, Kullback-leibler approach to gaussian mixture reduc- tion, IEEE Transactions on Aerospace and Electronic Systems 43 (3) (2007) 989–999

    2007

  39. [47]

    N. Wang, D. Forger, The level set kalman filter for state estimation of continuous-discrete systems, IEEE Transactions on Signal Processing (2021)

    2021

  40. [48]

    A. G. Wills, J. Hendriks, C. Renton, B. Ninness, A bayesian filtering algorithm for gaussian mixture models (2023).arXiv:1705.05495

    2023 arXiv

  41. [49]

    A. H. Stroud, Approximate calculation of multiple integrals, Prentice Hall, 1971

    1971

  42. [50]

    Cools, Constructing cubature formulae: the science behind the art, Acta numerica 6 (1997) 1–54

    R. Cools, Constructing cubature formulae: the science behind the art, Acta numerica 6 (1997) 1–54. 43

    1997

  43. [51]

    Y. Wang, H. Zhang, Accurate gaussian sum-filter for continuous-discrete nonlinear systems with non-gaussian noise, in: 2018 10th Interna- tional Conference on Communications, Circuits and Systems (ICCCAS), IEEE, 2018, pp. 140–145

    2018

  44. [52]

    Kandepu, B

    R. Kandepu, B. Foss, L. Imsland, Applying the unscented kalman filter for nonlinear state estimation, Journal of Process Control 18 (7) (2008) 753–768

    2008

  45. [53]

    H. Wu, H. Hu, S. Chen, G. Zeng, A gnss interference source tracking methodusingthecontinuous-discretegaussiankernelquadraturekalman filter, GPS Solutions 27 (4) (2023) 182

    2023

  46. [54]

    M. F. Emzir, Z. Zhao, S. Särkkä, Multidimensional projection filters via automatic differentiation and sparse-grid integration, Signal Processing 204 (2023) 108832

    2023

  47. [55]

    M. F. Emzir, Efficient projection filter algorithm for stochastic dynam- ical systems with correlated noises and state-dependent measurement covariance, Signal Processing 218 (2024) 109383

    2024

  48. [56]

    Njinkeu Nganso, S

    E. Njinkeu Nganso, S. G. Ngueuteu Mbouna, R. Yamapi, G. Filatrella, J. Kurths, White noise induces transitions between dynamical states in a network of nonlocally coupled birhythmic van der pol-type systems, Nonlinear Dynamics 113 (12) (2025) 15377–15391

    2025

  49. [57]

    Rodríguez, F

    M. Rodríguez, F. Blesa, R. Barrio, Opencl parallel integration of or- dinary differential equations: Applications in computational dynamics, Computer Physics Communications 192 (2015) 228–236

    2015

  50. [58]

    Schnoerr, G

    D. Schnoerr, G. Sanguinetti, R. Grima, Approximation and inference methods for stochastic biochemical kinetics—a tutorial review, Journal of Physics A: Mathematical and Theoretical 50 (9) (2017) 093001

    2017

  51. [59]

    J. K. Møller, H. Madsen, From state dependent diffusion to constant diffusion in stochastic differential equations by the lamperti transform (2010)

    2010

  52. [60]

    Bartosh, D

    G. Bartosh, D. Vetrov, C. A. Naesseth, Sde matching: Scalable and simulation-free training of latent stochastic differential equations, arXiv preprint arXiv:2502.02472 (2025). 44

    2025 arXiv

  53. [61]

    Rozet, G

    F. Rozet, G. Louppe, Score-based data assimilation, Advances in Neural Information Processing Systems 36 (2023) 40521–40541

    2023

  54. [62]

    F. Bao, Z. Zhang, G. Zhang, A score-based filter for nonlinear data assimilation, Journal of Computational Physics 514 (2024) 113207

    2024

  55. [63]

    E. Yoon, W. Chang, D. Kim, D. W. Kim, Rethinking forward processes for score-based nonlinear data assimilation in high dimensions, arXiv preprint arXiv:2604.02889 (2026)

    2026 arXiv

  56. [64]

    Mayer, D

    C. Mayer, D. W. Kim, M. Zhang, M. P. Lee, D. B. Forger, H. J. Burgess, C. Moon, Predicting circadian phase in community-dwelling later-life adults using actigraphy data, Journal of sleep research 34 (4) (2025) e14425

    2025

  57. [65]

    Huang, C

    Y. Huang, C. Mayer, P. Cheng, A. Siddula, H. J. Burgess, C. Drake, C. Goldstein, O. Walch, D. B. Forger, Predicting circadian phase across populations: a comparison of mathematical models and wearable de- vices, Sleep 44 (10) (2021) zsab126

    2021

  58. [66]

    M. P. Lee, D. W. Kim, Y. Fang, R. Kim, A. S. Bohnert, S. Sen, D. B. Forger, The real-world association between digital markers of circadian disruption and mental health risks, NPJ digital medicine 7 (1) (2024) 355

    2024

  59. [67]

    D. Lim, J. Jeong, Y. M. Song, C.-H. Cho, J. W. Yeom, T. Lee, J.-B. Lee, H.-J. Lee, J. K. Kim, Accurately predicting mood episodes in mood disorder patients using wearable sleep and circadian rhythm features, NPJ Digital Medicine 7 (1) (2024) 324

    2024

  60. [68]

    P. J. Van Leeuwen, Particle filtering in geophysical systems, Monthly Weather Review 137 (12) (2009) 4089–4114

    2009

  61. [69]

    M. W. Woolrich, S. Jbabdi, B. Patenaude, M. Chappell, S. Makni, T. Behrens, C. Beckmann, M. Jenkinson, S. M. Smith, Bayesian analysis of neuroimaging data in fsl, Neuroimage 45 (1) (2009) S173–S186

    2009

  62. [70]

    Y. Song, J. Sohl-Dickstein, D. P. Kingma, A. Kumar, S. Ermon, B. Poole, Score-based generative modeling through stochastic differen- tial equations, arXiv preprint arXiv:2011.13456 (2020). 45

    2020 arXiv

  63. [71]

    Lipman, R

    Y. Lipman, R. T. Chen, H. Ben-Hamu, M. Nickel, M. Le, Flow matching for generative modeling, arXiv preprint arXiv:2210.02747 (2022)

    2022 arXiv

  64. [72]

    Albergo, N

    M. Albergo, N. M. Boffi, E. Vanden-Eijnden, Stochastic interpolants: A unifyingframeworkforflowsanddiffusions, JournalofMachineLearning Research 26 (209) (2025) 1–80. 46 System ϵthreshold Wthreshold Cthreshold r Grid cell volume Van der Pol oscillator 0.05 10 −4 0.03 0.005 0.0...

    2025

  65. [73]

    ,2δ](Appendix B.2) = ∞X n=1 2 n−1 n! Dnv(x)[δ,

    The first term:Substituting2δinto (Appendix B.1) yields 1 2 v(x+ 2δ)−v(x) = 1 2 ∞X n=1 1 n! Dnv(x)[2δ, . . . ,2δ](Appendix B.2) = ∞X n=1 2 n−1 n! Dnv(x)[δ, . . . ,δ].(Appendix B.3)

  66. [74]

    ,δ].(Appendix B.4) Hence, the numerator in (13) is given by: 1 2 (v(x+ 2δ)−v(x)) −[v(x+δ)−v(x)] (Appendix B.5) = ∞X n=1 2n−1 −1 n! Dnv(x)[δ,

    The second term:(Appendix B.1) yields v(x+δ)−v(x) = ∞X n=1 1 n! Dnv(x)[δ, . . . ,δ].(Appendix B.4) Hence, the numerator in (13) is given by: 1 2 (v(x+ 2δ)−v(x)) −[v(x+δ)−v(x)] (Appendix B.5) = ∞X n=1 2n−1 −1 n! Dnv(x)[δ, . . . ,δ] .(Appendix B.6) Because 21−1−1 1! = 0, the lin...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.