pith. sign in

arxiv: 2606.18483 · v1 · pith:A5PIDYDXnew · submitted 2026-06-16 · 💻 cs.DC

Flexible Distributed Particle Filtering for the Internet of Things via Aggregate Computing

Angela Cortecchia , Davide Domini , Giovanni Ciatto , Roberto Casadei , Danilo Pianini , Mirko Viroli This is my paper

Pith reviewed 2026-06-26 22:16 UTC · model grok-4.3

classification 💻 cs.DC
keywords distributed particle filteringaggregate computingcomputational fieldsInternet of Thingsstate estimationsensor networkscyber-physical systems
0
0 comments X

The pith

Distributed particle filtering can be recast as computational fields to separate estimation logic from coordination strategies.

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

The paper proposes a field-based formulation of distributed particle filtering that uses aggregate computing to express both estimation and information dissemination. This decouples the core filtering operations from how data flows and coordination occurs across nodes. A sympathetic reader would care because most existing DPF solutions are locked to particular network architectures and communication patterns, which restricts their use in varied IoT settings. The formulation lets designers customize fusion-center placement, measurement aggregation functions, and the scope of information sharing inside one framework. In-silico experiments illustrate how different configurations trade off accuracy against communication cost and robustness.

Core claim

By expressing estimation and information dissemination as computational fields, our approach decouples the core filtering logic from coordination and data-flow strategies. This enables systematic customisation of key design dimensions, including fusion-center placement and resilience, aggregated measurement functions, as well as the type and scope of information propagation.

What carries the argument

Computational fields from aggregate computing that represent both the particle-filter estimation steps and the dissemination of measurements or particles.

If this is right

  • Diverse DPF configurations can be derived within a single unified framework.
  • Trade-offs among accuracy, communication cost, and robustness become directly comparable by varying field definitions.
  • The same abstraction supports adaptable DPF solutions across open and heterogeneous IoT environments.

Where Pith is reading between the lines

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

  • If the field abstraction works, it may let particle filtering be combined with other aggregate-computing services on the same IoT nodes without rewriting coordination code.
  • The approach could be tested by measuring how field evaluation overhead scales when the number of particles or network size increases on embedded devices.
  • Dynamic field reconfiguration might allow the filter to adapt its fusion strategy in response to changing network density or link quality.

Load-bearing premise

The assumption that recasting DPF operations as aggregate-computing fields preserves statistical correctness and convergence properties of the underlying particle filter.

What would settle it

Run the same set of particle-filter equations and network topology once with a conventional DPF implementation and once with the field-based version; the state estimates or particle weight distributions must match within sampling variance.

Figures

Figures reproduced from arXiv: 2606.18483 by Angela Cortecchia, Danilo Pianini, Davide Domini, Giovanni Ciatto, Mirko Viroli, Roberto Casadei.

Figure 1
Figure 1. Figure 1: Estimated trajectories for different numbers of sensors. With few neighbors (0-1), the filters fail to converge due to [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Average RMSE over 100 random seeds, with shaded [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Trajectories estimated via the leader-based fusion center [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
read the original abstract

State estimation from uncertain, distributed observations is central in many cyber-physical applications. While Distributed Particle Filtering (DPF) algorithms address nonlinear and non-Gaussian estimations in distributed settings, most solutions remain tied to specific architectures and communication assumptions, limiting adaptability in open, heterogeneous deployments-most notably, the Internet of Things (IoT). In this paper, we propose a field-based formulation of Distributed Particle Filtering grounded in Aggregate Computing (AC). By expressing estimation and information dissemination as computational fields, our approach decouples the core filtering logic from coordination and data-flow strategies. This enables systematic customisation of key design dimensions, including fusion-center placement and resilience, aggregated measurement functions, as well as the type and scope of information propagation. Through a set of in-silico experiments, we show how diverse DPF configurations can be derived within a unified framework, highlighting trade-offs among accuracy, communication cost, and robustness. Overall, the proposed approach positions AC as an effective abstraction layer for engineering adaptable DPF solutions in open IoT environments.

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 / 2 minor

Summary. The paper proposes a field-based formulation of Distributed Particle Filtering (DPF) grounded in Aggregate Computing (AC). By expressing estimation and information dissemination as computational fields, the approach decouples core filtering logic from coordination and data-flow strategies. This enables systematic customization of fusion-center placement, resilience, aggregated measurement functions, and information propagation. In-silico experiments demonstrate diverse DPF configurations within a unified framework and highlight trade-offs among accuracy, communication cost, and robustness, positioning AC as an abstraction layer for adaptable DPF in open IoT environments.

Significance. If the statistical equivalence to standard DPF holds, the work would offer a valuable abstraction for engineering flexible DPF solutions in heterogeneous IoT settings, where traditional DPF approaches are architecture-specific. The decoupling claim and experimental trade-off analysis could support reproducible design exploration, though the absence of an explicit invariance argument limits immediate impact.

major comments (2)
  1. [Abstract and Section 3 (field definitions)] The central claim requires that recasting DPF operations as AC fields leaves Bayesian updates, importance weighting, and resampling statistically unchanged. No equivalence argument or invariant is supplied showing that field operators produce identical per-particle quantities (weights, normalization, effective sample size) as centralized or gossip-based DPF; any deviation in aggregation or convergence would bias the posterior.
  2. [Section 4 and experimental evaluation] Section 4 (derived configurations) and the experimental section must demonstrate that the reported accuracy/communication/robustness trade-offs arise from the same underlying particle filter rather than from altered resampling or weighting semantics. Without this, the customization benefits cannot be isolated from potential statistical side-effects.
minor comments (2)
  1. [Abstract] The abstract refers to 'in-silico experiments' but provides no detail on particle count, network topology, or baseline DPF implementations used for comparison.
  2. [Section 2] Notation for field operators and their mapping to standard DPF steps (e.g., how aggregate measurement functions relate to likelihood evaluation) should be introduced earlier for clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address the two major comments point by point below, agreeing that an explicit statistical equivalence argument is needed to support the central claims.

read point-by-point responses

  1. Referee: [Abstract and Section 3 (field definitions)] The central claim requires that recasting DPF operations as AC fields leaves Bayesian updates, importance weighting, and resampling statistically unchanged. No equivalence argument or invariant is supplied showing that field operators produce identical per-particle quantities (weights, normalization, effective sample size) as centralized or gossip-based DPF; any deviation in aggregation or convergence would bias the posterior.

    Authors: We agree that the manuscript would benefit from an explicit invariance argument. The field-based formulation maps each DPF operation (prediction, measurement update, importance weighting, normalization, and resampling) directly onto field computations that are designed to produce identical per-particle quantities when fields converge correctly. However, we acknowledge the absence of a formal proof of statistical equivalence in the current version. We will add a new subsection in Section 3 that provides this argument, demonstrating that the field operators preserve the same weights, normalization constants, and effective sample size as standard DPF under the assumption of accurate field aggregation. revision: yes

  2. Referee: [Section 4 and experimental evaluation] Section 4 (derived configurations) and the experimental section must demonstrate that the reported accuracy/communication/robustness trade-offs arise from the same underlying particle filter rather than from altered resampling or weighting semantics. Without this, the customization benefits cannot be isolated from potential statistical side-effects.

    Authors: The reported trade-offs are intended to stem solely from variations in coordination and information propagation (expressed as fields), while the core filtering logic remains fixed. To make this explicit, we will revise Section 4 to include a clear statement that all configurations use identical prediction, update, weighting, and resampling steps. We will also add a short verification subsection in the experimental evaluation confirming that, for equivalent information availability, the particle weights and resulting posteriors match those produced by a centralized or gossip-based DPF implementation. revision: yes

Circularity Check

0 steps flagged

No circularity detected in derivation

full rationale

The manuscript presents a field-based reformulation of distributed particle filtering via aggregate computing as an engineering abstraction that decouples core filtering from coordination. No equations, parameter fits, or self-citations are exhibited that would reduce any claimed preservation of statistical properties, decoupling benefit, or derived configuration to the inputs by construction. The central premise is an unproven modeling assumption about invariance under field operators, which is a correctness gap rather than a definitional loop or self-referential prediction. The derivation chain therefore remains self-contained as a proposal of new design dimensions supported by experiments.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review; no free parameters, axioms, or invented entities are specified in the provided text.

pith-pipeline@v0.9.1-grok · 5722 in / 993 out tokens · 32939 ms · 2026-06-26T22:16:06.620364+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

28 extracted references

  1. [1]

    Distributed particle filtering in agent networks: A survey, classification, and comparison,

    O. Hlinka, F. Hlawatsch, and P. M. Djuric, “Distributed particle filtering in agent networks: A survey, classification, and comparison,”IEEE Signal Process. Mag., vol. 30, no. 1, pp. 61–81, 2013

    2013

  2. [2]

    A tutorial on particle filters for online nonlinear/non-gaussian bayesian tracking,

    M. S. Arulampalam, S. Maskell, N. J. Gordon, and T. Clapp, “A tutorial on particle filters for online nonlinear/non-gaussian bayesian tracking,” IEEE Trans. Signal Process., vol. 50, no. 2, pp. 174–188, 2002

    2002

  3. [3]

    Aggregate programming for the internet of things,

    J. Beal, D. Pianini, and M. Viroli, “Aggregate programming for the internet of things,”Computer, vol. 48, no. 9, pp. 22–30, 2015

    2015

  4. [4]

    Macroprogramming: Concepts, state of the art, and op- portunities of macroscopic behaviour modelling,

    R. Casadei, “Macroprogramming: Concepts, state of the art, and op- portunities of macroscopic behaviour modelling,”ACM Comput. Surv., vol. 55, no. 13s, pp. 275:1–275:37, 2023

    2023

  5. [5]

    A higher- order calculus of computational fields,

    G. Audrito, M. Viroli, F. Damiani, D. Pianini, and J. Beal, “A higher- order calculus of computational fields,”ACM Trans. Comput. Log., vol. 20, no. 1, pp. 5:1–5:55, 2019

    2019

  6. [6]

    Co-fields: A physically inspired approach to motion coordination,

    M. Mamei, F. Zambonelli, and L. Leonardi, “Co-fields: A physically inspired approach to motion coordination,”IEEE Pervasive Comput., vol. 3, no. 2, pp. 52–61, 2004

    2004

  7. [7]

    A new approach to linear filtering and prediction problems,

    R. E. Kalman, “A new approach to linear filtering and prediction problems,”J. Basic Eng., vol. 82, no. 1, p. 35–45, 1960

    1960

  8. [8]

    From distributed coordination to field calculus and aggregate comput- ing,

    M. Viroli, J. Beal, F. Damiani, G. Audrito, R. Casadei, and D. Pianini, “From distributed coordination to field calculus and aggregate comput- ing,”J. Log. Algebraic Methods Program., vol. 109, 2019

    2019

  9. [9]

    Computation against a neighbour: Addressing large-scale distribution and adaptivity with functional programming and scala,

    G. Audrito, R. Casadei, F. Damiani, and M. Viroli, “Computation against a neighbour: Addressing large-scale distribution and adaptivity with functional programming and scala,”Log. Methods Comput. Sci., vol. 19, no. 1, 2023

    2023

  10. [10]

    Engineering resilient collective adaptive systems by self-stabilisation,

    M. Viroli, G. Audrito, J. Beal, F. Damiani, and D. Pianini, “Engineering resilient collective adaptive systems by self-stabilisation,”ACM Trans. Model. Comput. Simul., vol. 28, no. 2, pp. 16:1–16:28, 2018

    2018

  11. [11]

    Designing self-organising emergent systems based on information flows and feedback-loops,

    T. D. Wolf and T. Holvoet, “Designing self-organising emergent systems based on information flows and feedback-loops,” in1st International Conference on Self-Adaptive and Self-Organizing Systems, SASO. IEEE Computer Society, 2007, pp. 295–298

    2007

  12. [12]

    Self-stabilising priority-based multi-leader election and network partitioning,

    D. Pianini, R. Casadei, and M. Viroli, “Self-stabilising priority-based multi-leader election and network partitioning,” inIEEE International Conference on Autonomic Computing and Self-Organizing Systems, ACSOS’22. IEEE, 2022, pp. 81–90

    2022

  13. [13]

    Self-stabilizing systems in spite of distributed control,

    E. W. Dijkstra, “Self-stabilizing systems in spite of distributed control,” Commun. ACM, vol. 17, no. 11, pp. 643–644, 1974

    1974

  14. [14]

    The exchange calculus (XC): A functional programming language design for distributed collective systems,

    G. Audrito, R. Casadei, F. Damiani, G. Salvaneschi, and M. Viroli, “The exchange calculus (XC): A functional programming language design for distributed collective systems,”J. Syst. Softw., vol. 210, p. 111976, 2024

    2024

  15. [15]

    Protelis: practical aggregate program- ming,

    D. Pianini, M. Viroli, and J. Beal, “Protelis: practical aggregate program- ming,” inProceedings of the 30th Annual ACM Symposium on Applied Computing. ACM, 2015, pp. 1846–1853

    2015

  16. [16]

    FCPP to aggregate them all,

    G. Audrito and G. Torta, “FCPP to aggregate them all,”Sci. Comput. Program., vol. 231, p. 103026, 2024

    2024

  17. [17]

    Multiplatform self-organizing systems through a Kotlin- MP implementation of aggregate computing,

    A. Cortecchia, “Multiplatform self-organizing systems through a Kotlin- MP implementation of aggregate computing,” inIEEE International Conference on Autonomic Computing and Self-Organizing Systems, ACSOS’2024 - Companion. IEEE, 2024, pp. 155–157

    2024

  18. [18]

    Distributed Kalman filter with embedded consensus filters,

    R. Olfati-Saber, “Distributed Kalman filter with embedded consensus filters,” in44th IEEE Conference on Decision and Control & 8th European Control Conference Control. IEEE, 2005, pp. 8179–8184

    2005

  19. [19]

    Diffusion strategies for distributed kalman filtering and smoothing,

    F. S. Cattivelli and A. H. Sayed, “Diffusion strategies for distributed kalman filtering and smoothing,”IEEE Trans. Autom. Control., vol. 55, no. 9, pp. 2069–2084, 2010

    2069

  20. [20]

    Likeli- hood consensus and its application to distributed particle filtering,

    O. Hlinka, O. Sluciak, F. Hlawatsch, P. M. Djuric, and M. Rupp, “Likeli- hood consensus and its application to distributed particle filtering,”IEEE Trans. Signal Process., vol. 60, no. 8, pp. 4334–4349, 2012

    2012

  21. [21]

    Consensus-based distributed particle filtering with distributed proposal adaptation,

    O. Hlinka, F. Hlawatsch, and P. M. Djuric, “Consensus-based distributed particle filtering with distributed proposal adaptation,”IEEE Trans. Signal Process., vol. 62, no. 12, pp. 3029–3041, 2014

    2014

  22. [22]

    Near-optimal knowledge- free resilient leader election,

    Y . Mo, G. Audrito, S. Dasgupta, and J. Beal, “Near-optimal knowledge- free resilient leader election,”Autom., vol. 146, p. 110583, 2022

    2022

  23. [23]

    Decentralized sensor fusion with distributed particle filters,

    M. Rosencrantz, G. J. Gordon, and S. Thrun, “Decentralized sensor fusion with distributed particle filters,” in19th Conference in Uncertainty in Artificial Intelligence. Morgan Kaufmann, 2003, pp. 493–500

    2003

  24. [24]

    Distributed particle filters for sensor networks,

    M. Coates, “Distributed particle filters for sensor networks,” in3rd International Symposium on Information Processing in Sensor Networks, IPSN’04, Proceedings. ACM, 2004, pp. 99–107

    2004

  25. [25]

    Cooperative parameter estimation on the unit sphere using a network of diffusion particle filters,

    C. G. de Figueredo, C. J. B. Jr., and M. G. S. Bruno, “Cooperative parameter estimation on the unit sphere using a network of diffusion particle filters,”IEEE Signal Process. Lett., vol. 27, pp. 715–719, 2020

    2020

  26. [26]

    Partitioned integration and coordination via the self-organising coordination regions pattern,

    D. Pianini, R. Casadei, M. Viroli, and A. Natali, “Partitioned integration and coordination via the self-organising coordination regions pattern,” Future Gener. Comput. Syst., vol. 114, pp. 44–68, 2021

    2021

  27. [27]

    Space- fluid adaptive sampling by self-organisation,

    R. Casadei, S. Mariani, D. Pianini, M. Viroli, and F. Zambonelli, “Space- fluid adaptive sampling by self-organisation,”Log. Methods Comput. Sci., vol. 19, no. 4, 2023

    2023

  28. [28]

    Chemical-oriented simulation of computational systems with alchemist,

    D. Pianini, S. Montagna, and M. Viroli, “Chemical-oriented simulation of computational systems with alchemist,”Journal of Simulation, vol. 7, no. 3, pp. 202–215, Aug. 2013

    2013