Pith. sign in

REVIEW 3 major objections 5 minor 38 references

DIFNet: Decentralized Information Filtering Fusion Neural Network with Unknown Correlation in Sensor Measurement Noises

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Pith's one-line read A GRU learns decentralized fusion weights that match centralized information filtering without knowing measurement noise correlations.

desk verdict The central theorem in this paper does not hold for reduced-order local models: Appendix A silently uses T^† T = I for non-square T, and the consequence is that the claimed equivalence to centralized EIF is false for arbitrary topologies with cross-correlated, non-overlapping sensors. read the letter →

arxiv 2508.18854 v1 pith:EVFYKPQQ submitted 2025-08-26 eess.SP

classification eess.SP
keywords decentralizedsensornetworksinformationfiltercross-correlatedmeasurementnoiseunknowncorrelationmodeldistributionGRUtargettrackingmultisensorfusion
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 aims to show that a decentralized sensor network can fuse noisy measurements almost as well as a centralized filter even when the correlations between sensor noises are unknown. Its theoretical core is Proposition 1: when each sensor's measurement Jacobian has full row rank and the local state transformation has rank at least the measurement dimension, the decentralized extended information filter with cross-correlated measurement noises is algebraically equivalent to the centralized EIF, provided each sensor contributes through a specific fusion weight matrix. The proposed DIFNet is a GRU-based network that learns those fusion weights from each sensor's local information contribution and information matrix, so the network never needs to know the noise cross-covariances. In simulated linear and nonlinear target tracking with a common jammer-induced correlation, DIFNet beats the decentralized information filter that uses wrong noise parameters and comes close to the exact-model filter, including when the noise is time-varying. If the learning claim holds, unknown correlation can be handled by data instead of by explicit covariance modeling, while retaining the communication and scalability advantages of decentralized fusion.

What carries the argument

The load-bearing object is the fusion weight M_j^k (and its local analogue) from Proposition 1. It is the factor that converts the centralized information update into a sum over local sensors, carrying all cross-correlation information through the appropriate column block of R_k^{-1}. DIFNet replaces the algebraic formula with the output of a GRU whose input is the local information contribution and information matrix from each sensor's EKF; the GRU's hidden state gives it temporal memory for time-varying correlations. The internodal transformation matrices T_j^k define which sensors communicate and keep the local subspaces consistent.

What would settle it

Train DIFNet on the four-sensor topology and jammer setting used in the paper, then test on a second jammer with a different cross-correlation structure or on a network where two sensors share only one state; if the RMSE falls back to the mismatched-DIF level, the learned weights are not the algebraic M_j^k but a scenario-specific fit. In the linear case, directly comparing the GRU output to the closed-form M_j^k of equation (17) would settle it.

Watch

Extended reading notes

Core claim

The central discovery is that the entire effect of cross-correlated measurement noise on decentralized fusion can be compressed into one matrix per sensor, M_j^k = ∇h_k^T R_k^{-1}(*j) R_j^k (∇h_j^k†)^T, which multiplies the local information contribution in the global update. Proposition 1 proves that with full-row-rank measurement Jacobians and rank(T_j^k) ≥ n_j, inserting these weights into the belief update makes the decentralized extended information filter exactly reproduce the centralized EIF; the same weights, transformed through internodal matrices, give consistent local estimates at every node. DIFNet then claims the GRU can regress this weight matrix from the pair (i-hat_j^k, I-hat

Load-bearing premise

The load-bearing premise is that a GRU, fed only with local information contributions and information matrices computed by EKFs with deliberately wrong noise parameters, can learn fusion weights that match the algebraic M_j^k and generalize to unseen correlations, noise patterns, and network topologies.

Editorial extensions

If this is right

  • If the learned weights reproduce Proposition 1's M_j^k, decentralized networks with arbitrary topologies and overlapping subspaces can match centralized EIF accuracy without a fully connected communication graph.
  • Sensors would not need to exchange raw measurements or estimate the cross-covariance sub-blocks R_ij^k; local information pairs suffice for fusion.
  • The model-distribution framework extends from linear systems to nonlinear extended information filters under cross-correlated measurement noises.
  • Under unknown time-varying noise, the learned filter can stay near exact-model performance while model-based filters with mismatched parameters degrade.
  • Distributed or federated training is compatible with the architecture, so privacy and scalability can be preserved while learning fusion rules.

Reading between the lines

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

  • An explicit comparison of the GRU's learned weights against the closed-form M_j^k in the linear case would reveal whether the network has learned the algebraic mapping or is compensating in a scenario-specific way; the paper does not show this.
  • The model-distribution proof shows exact equivalence is impossible when rank(T_j^k) < n_j, but DIFNet's learned weights might partially compensate even there; the paper does not test that regime.
  • Because the network uses temporal memory, the time-varying noise result suggests it may be learning a correlation function over time; varying the correlation pattern online would be a natural stress test.
  • The same weight-learning idea could transfer to track-to-track fusion or covariance-intersection settings where cross-correlations are also unknown, though DIFNet's current proof is tied to the information-filter update structure.
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

3 major / 5 minor

Summary. The paper proposes DIFNet, a GRU-based learning scheme for decentralized information filtering with unknown cross-correlated measurement noises. The model-based component derives a decentralized extended information filter (DEIF) with cross-correlated noises (Eqs. 16–19) and claims equivalence to the centralized EIF under full-row-rank measurement Jacobians and rank(T_j) ≥ n_j. The data-driven component uses local information contributions (i_hat_j, I_hat_j) as inputs to a GRU network that estimates the fusion weights M_j^k, trained by MSE against ground-truth state trajectories. Experiments on a single four-sensor topology compare DIFNet with DIF-exact, DIF-inexact, and centralized KF/EKF in linear, nonlinear, and time-varying-noise settings.

Significance. If the local equivalence in Proposition 1 were correct, the paper would offer a principled route to decentralized learning of cross-correlation structure with reduced communication, and the public source code is a definite strength. The global update (16) is a useful decomposition and the DIF-exact/centralized agreement is a good sanity check. However, the local subspace claim is algebraically false for the reduced-order local models that the paper explicitly targets, so the theoretical foundation is unsound. The empirical evidence is also thin: single-seed RMSE curves, one fixed topology, no error bars, and a supervised regression protocol trained and tested on the same simulator distribution. The claimed 'arbitrary network topologies' and 'learning unknown correlations' are not established by the present manuscript.

major comments (3)
  1. [Appendix A, Eqs. (A.4a)–(A.4b); Proposition 1 (Sec. 2.2.3)] The local subspace equivalence is derived by rewriting (T_j)^T I_hat_j T_j T_i^† as (T_i)^T (T^{ij})^T I_hat_j T^{ij}. This step requires T^{ij} T_i = T_j, i.e., T_i^† T_i = I_m. For reduced-order local models with m_i < m, T_i^† T_i is only the orthogonal projector onto row(T_i), not the identity. Minimal counterexample: m=2, T_1=[1,0], T_2=[0,1], H=I, R_11=R_22=1, R_12=0.5, P_pred=I, z_1=1, z_2=2. The centralized posterior for state 1 is 4/15, while Eq. (18a) at node 1 gives (4/3)/(7/3)=4/7. Thus Proposition 1's claim that the local estimates are equivalent to centralized EIF is false unless T_i are square/full-column-rank or cross-correlations are restricted to overlapping subspaces. Remark 3 only discusses rank(T_j)<n_j and does not cover this case. The experiments use T_1=[I_4,0] and T_4=[0,I_2], so the theorem as stated does not apply to them; the reported DIF-exact/centralized agr
  2. [Sec. 3.2 and Sec. 4] The central learnability assumption is unverified. The paper asserts that the pair (i_hat_j, I_hat_j) computed from local EKFs with deliberately wrong noise parameters is a sufficient input for a GRU to regress the optimal fusion weights M_j^k, but no representational analysis or ablation is provided. Only one network topology (Fig. 4) is tested, and all RMSE curves appear to come from a single training/evaluation run without error bars. Claims such as 'DIFNet enables consistent estimation in decentralized systems with arbitrary network topologies' are therefore not supported by the evidence. At minimum, the authors should report Monte Carlo runs with error bars, test on unseen cross-correlation patterns and at least one additional topology, and compare against the IFNet baseline and a covariance-intersection baseline.
  3. [Sec. 4.2–4.4] The numerical protocol is a supervised interpolation test rather than a prediction test. DIFNet is trained on ground-truth-labeled trajectories generated from the same simulator with the same noise statistics and topology used in testing. The improvement over DIF-inexact is therefore expected and does not demonstrate generalization to unknown correlations or network geometries. The time-varying experiments only vary the scalar sigma within the same structure. The paper should acknowledge this limitation and temper the conclusion that DIFNet 'learns unknown noise correlations' in a model-free sense.
minor comments (5)
  1. [Throughout] Numerous typos and grammatical errors: 'Teh', 'sensro', 'comminication', 'Nolinear', 'thar', 'informaiton', 'T able' in Table 3. A careful proofread is needed.
  2. [Sec. 4, RMSE definition] The RMSE formula uses N both for the number of sensors and for the number of independent runs. Please disambiguate these quantities.
  3. [Eqs. (17) and (19)] The notation R_k(∗j)^{-1} R_j is confusing; it should be written as R^{-1}_k(∗j) R_j to indicate the j-th column block of R^{-1}_k.
  4. [Sec. 4.3] The list of sensor positions includes s0, s1, s2, s4 but no s3; this appears to be a typo.
  5. [Sec. 4.5] The time-consumption table compares only DIF exact and DIFNet; no centralized or DIF-inexact baselines are included, and raw times are not reported. The conclusion that DIFNet's cost is 'acceptable' would be stronger with these details.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the model-based fusion identity is an algebraic reformulation and the learning results are standard supervised regression evaluated on held-out test data.

full rationale

The central derivation (Proposition 1) rewrites the centralized EIF update (6) by splitting R^{-1} into column blocks and inserting R_j(∇h_j^†)^T(∇h_j)^T R_j^{-1} = I, which is valid under the stated full-row-rank assumption on ∇h_j. This is an exact algebraic decomposition, not a fit or a self-referential definition. The local equations (18-19) are obtained by projecting the global update with (T_i^†)^T and T_i^†, so the claimed consistency with the global filter is a direct algebraic consequence, independent of any learned parameter. The DIFNet component minimizes the MSE loss (22) against ground-truth labels on a training set of 100 trajectories and is evaluated on a separate test set of 40 trajectories; thus the reported RMSE is a held-out generalization measurement, not a fitted quantity renamed as a prediction. The only self-citation is the use of the input-output configuration 'validated for its effectiveness in [13]' (Section 3.2), but this choice is also motivated by the exact information contributions appearing in Proposition 1, and the paper's own experiments provide independent evidence, so the citation is not load-bearing. Any algebraic concern about T_i^† T_i ≠ I for non-square T_i in Appendix A is a mathematical correctness issue, not a circularity, and is therefore not scored here.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The model-based fusion proof introduces no new physical entities; it relies on standard rank assumptions and pseudo-inverse algebra. The learning component depends on hand-set NN hyperparameters and the unproven sufficiency of local information contributions as network inputs. The numerical claims are tied to one simulated jammer scenario and one network topology.

free parameters (4)
  • Neural network architecture hyperparameters = unspecified (GRU hidden size is 'an integer product of m^2'; FC layer sizes not given)
    Architecture capacity directly determines whether the fusion-weight mapping can be represented; without exact values, reproduction is incomplete.
  • Loss regularization coefficient gamma = not reported
    Equation (22) includes gamma * ||Theta||^2; its value affects training and is not reported.
  • Optimizer and scheduler settings = Adam lr=1e-3, batch=20, weight decay=1e-4, CyclicLR settings unreported
    Training details are hand-set and not justified; final performance depends on them.
  • Simulation constants for the jammer scenario = beta_j=0.5 (linear), beta_j=2 (nonlinear), sigma=0.5, q=5 for inexact baselines
    These parameters define the test distribution and the difficulty of the baseline; results are specific to them.
assumptions (6)
  • domain assumption Measurement Jacobians grad h_j^k have full row rank and rank(T_j^k) >= n_j
    Proposition 1's equivalence to centralized EIF relies on this; it is stated after the proposition and in Remark 3.
  • domain assumption Measurement noises are white Gaussian and cross-correlated only within the same time instant
    The state-space model (1)-(3) assumes this; temporal correlation of measurement noise is not treated.
  • domain assumption EKF linearization errors are negligible in the nonlinear experiments
    The proof uses Jacobians grad f and grad h; the paper does not analyze the effect of linearization error on DIFNet.
  • ad hoc to paper The pair (i-hat_j^k, I-hat_j^k) is a sufficient input feature set for learning optimal fusion weights M_j^k
    Section 3.2 adopts IFNet's input-output configuration without proof that the mapping is learnable from these features, especially when local filter parameters are wrong.
  • domain assumption Network connectivity is determined by overlapping information spaces via T_ij^k
    Remark 2 defines communication this way; experiments only exercise one connectivity pattern.
  • standard math Moore-Penrose pseudo-inverse identities for full-row-rank matrices
    Appendix A uses (grad h_j^dagger)^T (grad h_j)^T = I; this is standard linear algebra from [38].

how reviews work

0 comments
Cite this review

Pith. "Pith review of DIFNet: Decentralized Information Filtering Fusion Neural Network with Unknown Correlation in Sensor Measurement Noises." pith.science (2026). https://pith.science/paper/EVFYKPQQ

@misc{pith2026250818854,
  author       = {Pith},
  title        = {Pith review of: DIFNet: Decentralized Information Filtering Fusion Neural Network with Unknown Correlation in Sensor Measurement Noises},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EVFYKPQQ}},
  note         = {Machine review of arXiv:2508.18854}
}
read the original abstract

In recent years, decentralized sensor networks have garnered significant attention in the field of state estimation owing to enhanced robustness, scalability, and fault tolerance. Optimal fusion performance can be achieved under fully connected communication and known noise correlation structures. To mitigate communication overhead, the global state estimation problem is decomposed into local subproblems through structured observation model. This ensures that even when the communication network is not fully connected, each sensor can achieve locally optimal estimates of its observable state components. To address the degradation of fusion accuracy induced by unknown correlations in measurement noise, this paper proposes a data-driven method, termed Decentralized Information Filter Neural Network (DIFNet), to learn unknown noise correlations in data for discrete-time nonlinear state space models with cross-correlated measurement noises. Numerical simulations demonstrate that DIFNet achieves superior fusion performance compared to conventional filtering methods and exhibits robust characteristics in more complex scenarios, such as the presence of time-varying noise. The source code used in our numerical experiment can be found online at https://wisdom-estimation.github.io/DIFNet_Demonstrate/.

Figures

Figures reproduced from arXiv: 2508.18854 by the authors.

Figure 1
Figure 1. Fusion Architecture. The fusion center refers to the processing facilities. (a) Centralized: the fusion center computes the global estimate directly from all sensor measurements z. (b) Hierarchical: low-level fusion centers compute local estimates xˆ, Pˆ from their own sensor measurements z, and a high-level fusion center combines these local estimates into a global estimate. (c) Decentralized: each fusion center co… view at source ↗
Figure 2
Figure 2. Schematic diagram of the proposed DIFNet. The top part shows the generation of local estimates and error covariances by local sensors. The bottom part illustrates the communication and implementation flow at each fusion center. require explicit cross-correlation modeling, our approach learns these relationships implicitly through neural net￾works. This federated learning inspiration is reflected in 375 three core pr… view at source ↗
Figure 3
Figure 3. RNN block of DIFNet in each sensor, consisting of FC and GRU. put of its neural network ˆi i k and ˆI i k locally, and the fusion weights Mij k 420 are estimated using the neural network. Fi￾nally, based on the obtained fusion weights, the received information is fused with the local state estimation to ob￾tain the final state estimation. For the training algorithm, let a dataset D =  (Xl , Z 1,l , · · · , Z N,l) L… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: The comminication relationship of 4 sensors in sensor network. We assume that the natural measurement noise of the 4 sensors w j k , for j=1,2,3,4, are mutually independent. How￾ever there is a jammer signal w0 k with covariance R0 495 k which 11 [PITH_FULL_IMAGE:figu…
Figure 5
Figure 5. Figure 5: RMSE of estimated position under linear state space model, tested on the test dataset. 0 10 20 30 40 50 time (s) 2 4 6 8 10 12 RMSE of velocity (m/s) sensor 1 DIF_exact sensor 1 DIF_inexact Centralized KF sensor 1 Decentralized IFNet 0 10 20 30 40 50 time (s) 2 4 6 8 1…
Figure 6
Figure 6. Figure 6: RMSE of estimated velocity under linear state space model, tested on the test dataset 4.3. Nolinear state space model In this experiment, we consider a target with state [x, x, y, ˙ y, z, ˙ z˙] T which has a constant speed and ˙x and ˙y have a constant known turn rate.…
Figure 8
Figure 8. Figure 8: RMSE of estimated velocity under nolinear state space model, tested on the test dataset covariance matrix Rvar k under time-varying case at time instance k satisfies the following equation Rvar k = (1 + σ cos(2π k T )R˜ k, where R˜ 580 k which is unknown to the sensor …
Figure 10
Figure 10. Figure 10: Fig.10 [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]
Figure 7
Figure 7. Figure 7: RMSE of estimated position under nolinear state space model, tested on the test dataset. 4.4. Time-varying noise In practice, noise not only has unknown correlations, 575 but it may also be time-varying. In this part, we will test DIFNet under time-varying measurement …
Figure 9
Figure 9. Figure 9: RMSE of estimated position under time-varying (σ = 0.5) measurement noise case, tested on the test dataset. From Fig.9 and Fig.10, we can observe that DIFNet can 14 [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: RMSE of estimated velocity under time-varying (σ = 0.5) measurement noise case, tested on the test dataset. still achieve similar results to DIF with precise model pa￾rameters (except σ), and performs better in estimating 590 velocity than in the linear state space mo…
Figure 11
Figure 11. Figure 11: RMSE of estimated position under different values of σ, tested on the test dataset. From Fig.11 and Fig.12, we can observe that under dif￾0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 4 5 6 7 8 9 10 RMSE of velocity (m/s) sensor 1 DIF_exact sensor 1 DIF_inexact Centralized KF s…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 37 canonical work pages

  1. [1]

    Bar-Shalom, P

    Y. Bar-Shalom, P. K. Willett, X. Tian, Tracking and data fu- sion, Vol. 11, YBS publishing Storrs, CT, USA:, 2011

  2. [2]

    Chong, Forty years of distributed estimation: A review of noteworthy developments, 2017 Sensor Data Fusion: Trends,705 Solutions, Applications (SDF) (2017) 1–10

    C.-Y. Chong, Forty years of distributed estimation: A review of noteworthy developments, 2017 Sensor Data Fusion: Trends,705 Solutions, Applications (SDF) (2017) 1–10

  3. [3]

    Bar-Shalom, X

    Y. Bar-Shalom, X. R. Li, Estimation and tracking: Principles, techniques, and software, Artech House (1993)

  4. [4]

    B. Yang, B. Chen, Y. Liu, J. Chen, Gaussian process fusion method for multi-fidelity data with heterogeneity distribution710 in aerospace vehicle flight dynamics, Engineering Applications of Artificial Intelligence 138 (2024) 109228

  5. [5]

    Kheirandish, E

    M. Kheirandish, E. A. Yazdi, H. Mohammadi, M. Mohammadi, A fault-tolerant sensor fusion in mobile robots using multiple model kalman filters, Robotics and autonomous systems 161715 (2023) 104343

  6. [6]

    Huang, S

    Z. Huang, S. Sun, J. Zhao, L. Mao, Multi-modal policy fu- sion for end-to-end autonomous driving, Information Fusion 98 (2023) 101834

  7. [7]

    Hall, C.-Y

    D. Hall, C.-Y. Chong, J. Llinas, M. Liggins II, Distributed data720 fusion for network-centric operations, Crc Press, 2017

  8. [8]

    Chong, K.-C

    C.-Y. Chong, K.-C. Chang, S. Mori, A review of forty years of distributed estimation, in: 2018 21st International Conference on Information Fusion (FUSION), IEEE, 2018, pp. 1–8

Show all 38 references
  1. [9]

    X. Li, Y. Zhu, J. Wang, C. Han, Optimal linear estimation fu-725 sion .I. Unified fusion rules, IEEE Transactions on Information Theory 49 (9) (2003) 2192–2208

  2. [10]

    Mutambara, H

    A. Mutambara, H. F. Durrant-Whyte, A formally verified mod- ular decentralized robot control system, in: Proceedings of 1993 IEEE/RSJ International Conference on Intelligent Robots and730 Systems (IROS’93), Vol. 3, IEEE, 1993, pp. 2023–2030

  3. [11]

    A. G. Mutambara, Decentralized Estimation and Control for Multisensor Systems, CRC press, 1998

  4. [12]

    Greenway, R

    P. Greenway, R. Deaves, D. Bull, Communications management in decentralised data fusion systems, in: 1996 IEEE/SICE/RSJ735 International Conference on Multisensor Fusion and Integration for Intelligent Systems (Cat. No. 96TH8242), IEEE, 1996, pp. 796–805

  5. [13]

    M. Wang, H. Liu, H. Tang, M. Zhang, X. Shen, IFNet: Data- driven multisensor estimate fusion with unknown correlation740 in sensor measurement noises, Information Fusion 115 (2025) 102750

  6. [14]

    Bar-Shalom, On the track-to-track correlation problem, IEEE Transactions on Automatic control 26 (2) (1981) 571– 572.745

    Y. Bar-Shalom, On the track-to-track correlation problem, IEEE Transactions on Automatic control 26 (2) (1981) 571– 572.745

  7. [15]

    Bar-Shalom, L

    Y. Bar-Shalom, L. Campo, The effect of the common process noise on the two-sensor fused-track covariance, IEEE Transac- tions on aerospace and electronic systems AES-22 (6) (1986) 803–805

  8. [16]

    S. Sun, Z. Deng, Multi-sensor optimal information fusion750 Kalman filter, Automatica 40 (6) (2004) 1017–1023

  9. [17]

    S. J. Julier, J. K. Uhlmann, A non-divergent estimation al- gorithm in the presence of unknown correlations, in: Pro- ceedings of the 1997 American Control Conference (Cat. No. 97CH36041), Vol. 4, IEEE, 1997, pp. 2369–2373.755

  10. [18]

    J. K. Uhlmann, General data fusion for estimates with unknown cross covariances, in: Signal Processing, Sensor Fusion, and Tar- get Recognition V, Vol. 2755, SPIE, 1996, pp. 536–547

  11. [19]

    Julier, J

    S. Julier, J. K. Uhlmann, General decentralized data fusion with covariance intersection, in: Handbook of multisensor data fu-760 sion, CRC Press, 2017, pp. 339–364

  12. [20]

    Revach, N

    G. Revach, N. Shlezinger, X. Ni, A. L. Escoriza, R. J. G. van Sloun, Y. C. Eldar, Kalmannet: Neural network aided kalman filtering for partially known dynamics, IEEE Transactions on Signal Processing 70 (2022) 1532–1547.765

  13. [21]

    Willner, C.-B

    D. Willner, C.-B. Chang, K.-P. Dunn, Kalman filter algorithms for a multi-sensor system, in: 1976 IEEE conference on de- cision and control including the 15th symposium on adaptive processes, IEEE, 1976, pp. 570–574

  14. [22]

    Hassan, G

    M. Hassan, G. Salut, M. Singh, A. Titli, A decentralized com-770 putational algorithm for the global kalman filter, IEEE Trans- actions on Automatic Control 23 (2) (1978) 262–268

  15. [23]

    H. R. Hashemipour, S. Roy, A. J. Laub, Decentralized struc- tures for parallel kalman filtering, IEEE Transactions on auto- matic control 33 (1) (1988) 88–94.775

  16. [24]

    B. S. Rao, H. F. Durrant-Whyte, Fully decentralised algorithm for multisensor kalman filtering, in: IEE Proceedings D (Control Theory and Applications), Vol. 138, IET, 1991, pp. 413–420

  17. [25]

    H. F. Durrant-Whyte, B. Rao, H. Hu, Toward a fully decentral- ized architecture for multi-sensor data fusion, in: Proceedings.,780 IEEE International Conference on Robotics and Automation, IEEE, 1990, pp. 1331–1336

  18. [26]

    S. Oruc, J. Sijs, P. Van den Bosch, Optimal decentralized kalman filter, in: 2009 17th Mediterranean Conference on Con- trol and Automation, IEEE, 2009, pp. 803–808.785

  19. [27]

    Chang, Y

    K.-C. Chang, Y. Bar-Shalom, Distributed adaptive estimation with probabilistic data association, Automatica 25 (3) (1989) 359–369

  20. [28]

    Chang, Y

    K.-C. Chang, Y. Bar-Shalom, Distributed multiple model es- timation, in: 1987 American Control Conference, IEEE, 1987,790 pp. 797–802

  21. [29]

    Chong, Distributed multitarget multisensor track- ing, Multitarget-multisensor tracking: Advanced applications (1990) 247–296

    C.-Y. Chong, Distributed multitarget multisensor track- ing, Multitarget-multisensor tracking: Advanced applications (1990) 247–296

  22. [30]

    Grime, H

    S. Grime, H. F. Durrant-Whyte, Data fusion in decentralized795 19 sensor networks, Control engineering practice 2 (5) (1994) 849– 863

  23. [31]

    Grime, H

    S. Grime, H. F. Durrant-Whyte, P. Ho, Communication in decentralized data-fusion systems, in: 1992 American Control Conference, IEEE, 1992, pp. 3299–3303.800

  24. [32]

    Utete, H

    S. Utete, H. F. Durrant-Whyte, Routing for reliability in de- centralised sensing networks, in: Proceedings of 1994 American Control Conference-ACC’94, Vol. 2, IEEE, 1994, pp. 2268–2272

  25. [33]

    Utete, H

    S. Utete, H. F. Durrant-Whyte, Reliability in decentralised data fusion networks, in: Proceedings of 1994 IEEE International805 Conference on MFI’94. Multisensor Fusion and Integration for Intelligent Systems, IEEE, 1994, pp. 215–221

  26. [34]

    T. M. Berg, H. F. Durrant-Whyte, Model distribution in decen- tralized multi-sensor data fusion, in: 1991 American Control Conference, IEEE, 1991, pp. 2292–2293.810

  27. [35]

    Chong, Hierarchical estimation, in: Proc

    C.-Y. Chong, Hierarchical estimation, in: Proc. MIT/ONR Workshop on C3, 1979, pp. 205–220

  28. [36]

    Y. Zhu, J. Zhou, X. Shen, E. Song, Y. Luo, Networked multi- sensor decision and estimation fusion: based on advanced math- ematical methods, CRC Press, 2012.815

  29. [37]

    X. Chen, Z. Sun, Y. Tao, S. S.-T. Yau, A uniform framework of yau–yau algorithm based on deep learning with the capability of overcoming the curse of dimensionality, IEEE Transactions on Automatic Control 70 (1) (2025) 339–354. doi:10.1109/TAC. 2024.3424628.820

  30. [38]

    Golub, W

    G. Golub, W. Kahan, Calculating the singular values and pseudo-inverse of a matrix, Journal of the Society for Indus- trial and Applied Mathematics, Series B: Numerical Analysis 2 (2) (1965) 205–224. 20

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.