Pith. sign in

REVIEW 2 major objections 5 minor 55 references

Confidence Boosts Trust-Based Resilience in Cooperative Multi-Robot Systems

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A confidence parameter that decays to zero lets a multi-robot consensus protocol tolerate any number of malicious agents while staying provably close to the adversary-free consensus.

desk verdict A genuinely useful refinement of trust-based consensus with a real speed/accuracy knob, but the main deviation bound has a repairable expectation-exchange gap in Appendix C. read the letter →

arxiv 2506.08807 v1 pith:5CNXHECI submitted 2025-06-10 eess.SP cs.MAcs.ROcs.SYeess.SY

Luca Ballotta , Áron Vékássy , Stephanie Gil , Michal Yemini This is my paper
classification eess.SPcs.MAcs.ROcs.SYeess.SY
keywords resilientconsensusmulti-robotsystemstrust-basedresiliencephysical-layertrustconfidenceparameterFriedkin-Johnsenmodelunderadversariesvehicularplatooning
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

The paper proposes a consensus update in which each robot's next state is a blend of its own initial state and a weighted average of trusted neighbors' states: $x_i(t+1)=\lambda_t x_i(0)+(1-\lambda_t)\sum_{j\in N_i\cup\{i\}} w_{ij}(t)x_j(t)$. The weights $w_{ij}(t)$ are built from physical-channel trust observations, and $\lambda_t$ is a confidence parameter that decays to zero. The main claim is that if trust observations are informative on average, this rule drives all legitimate robots to a common value almost surely, and the distance from the no-attack consensus is probabilistically bounded by an explicit expression involving the final misclassification times. The protocol needs no observation window before consensus starts and no prior bound on the number of malicious robots. This matters for applications such as platooning, where a designer can choose how fast $\lambda_t$ decays to trade convergence speed against closeness to the nominal consensus.

What carries the argument

The load-bearing object is the pair $(\beta_{ij}(t),\lambda_t)$: the aggregate trust score $\beta_{ij}(t)=\sum_{s=0}^t(\alpha_{ij}(s)-1/2)$ decides whether neighbor $j$ is trusted at time $t$, and the confidence parameter $\lambda_t=c e^{-\gamma t}$ decides how much the robot relies on its own initial state rather than on the trusted-neighbor average. The proof decomposes the trajectory into a legitimate contribution and a malicious contribution using the transition matrices $W^L_{t,\mathrm{aut}}$, $W^L_{t,\mathrm{in}}$, and $W^M_{k,t}$; it then invokes the almost-sure finiteness of the final classification time $T_f$ to replace learned weights by true weights after $T_f$, and uses a Jordan decomposition of the nominal weight matrix $W^L$ to quantify convergence. The exponential decay of misclassification probability from Lemma 1 is what makes the eventual classification argument work.

What would settle it

Run the platoon-merging experiment from Section VI for many Monte Carlo trials, estimate $\mathbb{E}[T_f]$ and $\mathbb{E}[T_M]$ from the Beta-distributed trust observations, compute $u_L+u_M$, and compare the empirical frequency of final deviations above $\epsilon$ with $(\eta/\epsilon)(u_L+u_M)$; a violation for a fixed $\epsilon$ would contradict Theorem 1. A sharper probe is to test a channel model with positively correlated trust observations across time (block fading) that still satisfies Assumption 2; if $\beta_{ij}(t)$ then fails to converge to the correct sign or $T_f$ is not almost surely finite, the independence implicit in Lemma 1 is the actual load-bearing assumption.

Watch

Extended reading notes

Core claim

The central discovery is that confidence, implemented as a decaying self-anchor $\lambda_t$, converts an uncertain trust signal into a formal resilience guarantee. Proposition 1 shows that under Assumptions 1 and 2 and $\lambda_t\to 0$, the legitimate robots reach consensus almost surely even with arbitrarily many malicious robots. Theorem 1 bounds the chance that the final consensus deviates from the no-attack value by more than $\epsilon$ as $\mathbb{P}[\lim_{t\to\infty}\tilde{x}_t^i>\epsilon]\le (\eta/\epsilon)(u_L+u_M)$, where $\eta$ bounds all states and $u_L$, $u_M$ are explicit functions of the expected final classification times $\mathbb{E}[T_f]$, $\mathbb{E}[T_M]$, the trust statistics $E_L$, $E_M$, the graph degrees, and the decay rate $\gamma$. The paper also derives a finite-time convergence-rate bound (Theorem 2) whose expectation is controlled by the tail of the misclassification time through $p(k)$. Numerical experiments with spoofed vehicles in merging platoons illustrate the speed-deviation tradeoff.

Load-bearing premise

The load-bearing premise is that trust scores are informative in expectation—legitimate transmissions average above $1/2$ and malicious below $1/2$—and that permanently correct classification of all neighbors happens in finite time almost surely; if the expected final classification time is enormous, the deviation guarantee becomes too weak to matter for any realistic mission.

Editorial extensions

If this is right

  • Legitimate robots can begin cooperating immediately at time zero; no synchronization or pre-agreed observation window is required.
  • For fixed channel statistics and network, the designer can tune $\gamma$ in $\lambda_t=c e^{-\gamma t}$ to move along the tradeoff: smaller $\gamma$ shrinks the expected deviation bound at the price of slower convergence, and larger $\gamma$ does the reverse.
  • The steady-state guarantee degrades as $\mathbb{E}[T_f]$ and $\mathbb{E}[T_M]$ grow, so slow-to-classify channels push the guaranteed accuracy down, while clearly identifiable malicious transmissions (large $|E_M|$) improve it.
  • After the final classification time, the protocol behaves approximately like standard consensus on the true legitimate weights, so the classical Perron-vector intuition applies to the final value.
  • The expected finite-time error bound in Theorem 2 gives a way to choose a deployment horizon based on the channel statistics.

Reading between the lines

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

  • The same proof structure should carry to vector-valued states and per-robot confidence schedules, since the state bound and weight matrices act componentwise; the constants would need recomputation.
  • One can view $\lambda_t$ as a regularization schedule in online distributed optimization: the deviation bound plays the role of optimization error and the convergence bound the role of convergence time, suggesting the tradeoff is fundamental to any trust-based scheme with uncertain classification.
  • A testable design rule suggested by the bounds is to choose $\gamma$ comparable to the misclassification decay rates $2E_L^2$ or $2E_M^2$, so the self-anchor fades on the same timescale as trust scores become reliable.
  • The bounds are Markov-inequality bounds and are therefore likely loose; the numerical tradeoff curves are flatter than the theory, meaning practical deployments may tolerate faster decay than the guarantees suggest.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 5 minor

Summary. This paper proposes a distributed resilient consensus protocol (RES) for multi-robot networks. Each legitimate robot updates its scalar state by combining an anchoring term λ_t x_i(0) with a weighted average of trusted neighbors' states, where λ_t is a time-varying confidence parameter chosen as λ_t = c e^{-γ t}. Trust observations from the physical layer are aggregated into scores β_ij(t), and communication weights are assigned to neighbors with nonnegative cumulative scores. The main analytical claims are: (i) Proposition 1: legitimate robots reach consensus almost surely for any vanishing λ_t; (ii) Propositions 2 and 3 bound the expected steady-state deviation from the nominal no-attack consensus, respectively due to legitimate and malicious contributions; (iii) Theorem 1 converts these into a probabilistic deviation bound P(limsup_t ˜x_i^t > ε) ≤ (η/ε)(u_L + u_M); (iv) Proposition 4 and Theorem 2 give finite-time convergence-rate bounds. Numerical experiments on vehicular platooning with spoofed vehicles illustrate the predicted speed-deviation tradeoff and compare with the step-window protocol of [14].

Significance. If the main theorem held, the paper would provide a valuable extension of trust-based resilient consensus: it removes the initial observation window of [14], handles uncertain channel-derived trust through a time-varying confidence parameter, and recovers exact consensus (not merely bounded disagreement) under a vanishing λ_t. The explicit bounds on deviation and convergence rate offer concrete design guidance, and the platooning simulations support the qualitative tradeoff. The appendices contain substantial original technical work, including Jordan-block estimates for the convergence rate and tail bounds for classification times. However, the headline quantitative guarantee rests on a single invalid expectation interchange in Appendix C; as written, Theorem 1 is not proven. The issue appears repairable by extending the relevant sum to infinity, but the resulting bound will differ from Eq. (36), so the proof needs to be redone before the results can be relied upon.

major comments (2)
  1. [Appendix C, Eq. (C.5)] The step E[η Σ_{k=0}^{TM(t)-1} a_k Y_k] = η Σ_{k=0}^{TM(t)-1} a_k E[Y_k], with a_k = (1-λ_{k+1})(1-λ_k) and Y_k = max_i [W^M_k 1]_i, is invalid because TM(t) is a random stopping index defined through the same trust-observation process that determines Y_k. The events {k < TM(t)} are not independent of Y_k, so the expectation of the product does not factor as written. This is the exact point at which the proof of Proposition 3, and hence the term u_M in Theorem 1, relies on an unproven interchange.
  2. [Appendix C, Eqs. (C.8)-(C.10), and Theorem 1, Eq. (37)] Because of the invalid interchange in Eq. (C.5), the subsequent computation E[ξ(TM)] ≤ D_M Σ_k ξ(k) e^{-2E_M^2 k} = D_M ζ in Eq. (C.10) does not follow, and the claimed bound u_M = D_M^2 ζ / 2 in Eq. (36) is not established. A correct and readily available route is to extend the sum to infinity and use Tonelli's theorem to justify E[Σ_{k=0}^∞ a_k Y_k] ≤ Σ_{k=0}^∞ a_k E[Y_k]; this produces a valid but different closed-form bound. Since the deviation bound (37) is the paper's headline quantitative guarantee, the proof of Proposition 3 and the statement of Theorem 1 must be revised.
minor comments (5)
  1. [Assumption 2 and Lemma 1] The exponential tail bounds in Eq. (5) require independence of the trust observations α_ij(t) across time and across links, but this independence assumption is not stated explicitly; please add it.
  2. [Appendix E, Eq. (E.13)] In the union-bound expression for P[T_f = k], the subscripts in the exponentials appear swapped: the term involving D_L should decay with E_L^2 and the term involving D_M should decay with E_M^2.
  3. [Proposition 4, Eqs. (40)-(44)] The notation bm is ambiguous: it is not clear whether this is a product b·m or a single constant. Please define the constant explicitly and write it as b m or with a subscript.
  4. [Appendix B, Eq. (B.8)] The definition of s(x) is typeset in a confusing way; Eq. (21) indicates the intended form s(x) = 1 + (1-x)/x ln(1-x), which should be stated directly.
  5. [Section VI] The description of the '2-nearest neighbor topology' should specify whether the communication graph is directed and how the cross-platoon leader links are chosen, since Assumption 1 depends on the structure of the resulting matrix W^L.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's analytical results are derived from first-principles assumptions and explicit bounds; self-citations provide independent supporting lemmas.

full rationale

The paper proves consensus (Proposition 1), deviation bounds (Propositions 2-3 and Theorem 1), and convergence rates (Proposition 4 and Theorem 2) directly from the proposed update rule (RES), the trust-based weights (4), and Assumptions 1-4. The quantities uL and uM are closed-form functions of the assumed channel statistics, classification times, and design constants, not fitted values or redefinitions of the target bounds. Cited prior results, including the Chernoff classification bounds from [14], the diminishing Friedkin-Johnsen identity from [16], and Lemma 2 of [43], are independent mathematical lemmas with assumptions that do not already contain the paper's consensus or deviation conclusions. Some cited works share authors, but they are not used as a self-referential justification of the main results. The proof step in Appendix C that exchanges an expectation with a sum whose random upper limit is TM(t) is a potential correctness gap, but it is an analysis error rather than a circular derivation, so it does not constitute circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The central claims rest on four standard but domain-specific assumptions: primitivity of the nominal weight matrix, statistical informativeness of physical-layer trust observations, bounded states, and vanishing λ_t. All are stated clearly. The only invented element is the confidence schedule itself, which is a design choice rather than a fitted parameter. The paper does not fit parameters to data: the only randomness is used probabilistically, and all bounds depend on quantities (E_L, E_M, T_f, T_M, d_M, γ) that are assumed or chosen rather than optimized against the simulation.

assumptions (4)
  • domain assumption Assumption 1: W_L is primitive; Perron eigenvector v exists.
    Basis for the nominal consensus value x_L,ss* = v^T x_0^L; used throughout Section V.
  • domain assumption Assumption 2: Trust observations are informative on average (E_L > 0, E_M < 0).
    Enters Lemma 1 on misclassification probabilities and all later bounds; no physical-channel model is given, so the validity is assumed without measurement evidence.
  • domain assumption Assumption 3: Bounded initial legitimate states and all malicious transmissions (η).
    Needed in Propositions 2 and 3; the paper justifies it by saying otherwise thresholding detects the adversary, but the thresholding is not part of the protocol.
  • ad hoc to paper Assumption 4: lim λ_t = 0.
    Required by Proposition 1 for consensus; the designer is assumed to pick such a schedule, but no autonomous method to choose the schedule is provided.
invented entities (1)
  • Confidence parameter λ_t with exponential schedule λ_t = c e^{-γ t}
    purpose: Controls how strongly each robot anchors to its initial state, enabling the speed-accuracy tradeoff.
    A design parameter, not an independently verified phenomenon. It is borrowed from the Friedkin-Johnsen competition literature, not a new physical or algorithmic entity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Confidence Boosts Trust-Based Resilience in Cooperative Multi-Robot Systems." pith.science (2026). https://pith.science/paper/5CNXHECI

@misc{pith2026250608807,
  author       = {Pith},
  title        = {Pith review of: Confidence Boosts Trust-Based Resilience in Cooperative Multi-Robot Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5CNXHECI}},
  note         = {Machine review of arXiv:2506.08807}
}
read the original abstract

Wireless communication-based multi-robot systems open the door to cyberattacks that can disrupt safety and performance of collaborative robots. The physical channel supporting inter-robot communication offers an attractive opportunity to decouple the detection of malicious robots from task-relevant data exchange between legitimate robots. Yet, trustworthiness indications coming from physical channels are uncertain and must be handled with this in mind. In this paper, we propose a resilient protocol for multi-robot operation wherein a parameter {\lambda}t accounts for how confident a robot is about the legitimacy of nearby robots that the physical channel indicates. Analytical results prove that our protocol achieves resilient coordination with arbitrarily many malicious robots under mild assumptions. Tuning {\lambda}t allows a designer to trade between near-optimal inter-robot coordination and quick task execution; see Fig. 1. This is a fundamental performance tradeoff and must be carefully evaluated based on the task at hand. The effectiveness of our approach is numerically verified with experiments involving platoons of autonomous cars where some vehicles are maliciously spoofed.

Figures

Figures reproduced from arXiv: 2506.08807 by the authors.

Figure 1
Figure 1. Our protocol allows robots to simultaneously cooperate and detect [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Upper bound u L in Proposition 2 on expected deviation due to legitimate robots for several values of Tf with dM = 9 and λt = 0.9e−γt . that a slow decay of λt reduces the deviation. This is also intuitive because, if λt decays slowly, the legitimate robots do not rely much on incoming messages for a long time, mitigating all misclassification. This reminds of the strategy in [14] where consensus starts at time T0 a… view at source ↗
Figure 3
Figure 3. Upper bound ρ(t) in Proposition 4 on convergence rate for a random geometric graph with L = 20 and Tf = TM = 50, and λt = 0.9e−γt . Proposition 4 (Convergence speed of (RES)). Let Assump￾tions 1 to 4 hold and Tf < ∞ be fixed. Define the coefficients D1 .= max i∈L |M ∩ Ni | |M ∩ Ni | + 1 , πt s .= Yt k=s (1 − λk). (40) Let σ be the second largest eigenvalue modulus of W L , mσ + 1 ≥ 1 the maximal size of Jordan block… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Mean distance from nominal consensus of legitimate vehicles. Parameter [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Mean distance from nominal consensus of legitimate vehicles. The [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

55 extracted references · 55 canonical work pages

  1. [14]

    Characterizing Trust and Resilience in Distributed Consensus for Cyberphysical Systems,

    M. Yemini, A. Nedi ´c, A. J. Goldsmith, and S. Gil, “Characterizing Trust and Resilience in Distributed Consensus for Cyberphysical Systems,” IEEE Trans. Robot. , vol. 38, no. 1, pp. 71–91, 2022

    work page 2022

  2. [16]

    Friedkin-Johnsen Model With Diminishing Competition,

    L. Ballotta, Á. Vékássy, S. Gil, and M. Yemini, “Friedkin-Johnsen Model With Diminishing Competition,” IEEE Control Syst. Lett. , vol. 8, pp. 2679–2684, 2024

    work page 2024

  3. [1]

    Energy-efficient communication for UA V-enabled mobile relay networks,

    ˙I. Ba¸ stürk, “Energy-efficient communication for UA V-enabled mobile relay networks,” Computer Networks, vol. 213, p. 109071, 2022

    work page 2022

  4. [2]

    Multirobot Coordination for Space Exploration,

    L. Yliniemi, A. K. Agogino, and K. Tumer, “Multirobot Coordination for Space Exploration,” AI Mag., vol. 35, no. 4, pp. 61–74, 2014

    work page 2014

  5. [3]

    Robotized and Automated Warehouse Systems: Review and Recent Developments,

    K. Azadeh, R. De Koster, and D. Roy, “Robotized and Automated Warehouse Systems: Review and Recent Developments,” Transp. Sci., vol. 53, no. 4, pp. 917–945, 2019

    work page 2019

  6. [4]

    A Novel MPC Formulation for Dynamic Target Tracking with Increased Area Coverage for Search-and- Rescue Robots,

    M. Baglioni and A. Jamshidnejad, “A Novel MPC Formulation for Dynamic Target Tracking with Increased Area Coverage for Search-and- Rescue Robots,” J. Intell. Robot. Syst. , vol. 110, no. 4, p. 140, 2024

    work page 2024

  7. [5]

    A Two- Stage Optimization-Based Motion Planner for Safe Urban Driving,

    F. Eiras, M. Hawasly, S. V . Albrecht, and S. Ramamoorthy, “A Two- Stage Optimization-Based Motion Planner for Safe Urban Driving,” IEEE Trans. Robot., vol. 38, no. 2, pp. 822–834, 2022

    work page 2022

  8. [6]

    Consensus-based fast and energy-efficient multi-robot task allocation,

    P. Mahato, S. Saha, C. Sarkar, and Md. Shaghil, “Consensus-based fast and energy-efficient multi-robot task allocation,” Robot. Auton. Syst. , vol. 159, p. 104270, 2023

    work page 2023

Show all 55 references
  1. [7]

    Byzantine Resilient Distributed Learning in Multirobot Systems,

    J. Li, W. Abbas, M. Shabbir, and X. Koutsoukos, “Byzantine Resilient Distributed Learning in Multirobot Systems,” IEEE Trans. Robot., vol. 38, no. 6, pp. 3550–3563, 2022

    2022

  2. [8]

    Decentralized rigidity maintenance control with range measurements for multi-robot systems,

    D. Zelazo, A. Franchi, H. H. Bülthoff, and P. Robuffo Giordano, “Decentralized rigidity maintenance control with range measurements for multi-robot systems,” Int. J. Robot. Res. , vol. 34, no. 1, pp. 105–128, 2015

    2015

  3. [9]

    Coordinated Control of Multi-Robot Systems: A Survey,

    J. Cortés and M. Egerstedt, “Coordinated Control of Multi-Robot Systems: A Survey,” SICE J. Control Meas. Syst. Integr. , vol. 10, no. 6, pp. 495– 503, 2017

    2017

  4. [10]

    Distributed Opti- mization Methods for Multi-Robot Systems: Part 2—A Survey,

    O. Shorinwa, T. Halsted, J. Yu, and M. Schwager, “Distributed Opti- mization Methods for Multi-Robot Systems: Part 2—A Survey,” IEEE Robot. Autom. Mag. , vol. 31, no. 3, pp. 154–169, 2024

    2024

  5. [11]

    Guaranteeing spoof-resilient multi-robot networks,

    S. Gil, S. Kumar, M. Mazumder, D. Katabi, and D. Rus, “Guaranteeing spoof-resilient multi-robot networks,” Auton. Robots, vol. 41, no. 6, pp. 1383–1400, 2017

    2017

  6. [12]

    State-Secrecy Codes for Networked Linear Systems,

    A. Tsiamis, K. Gatsis, and G. J. Pappas, “State-Secrecy Codes for Networked Linear Systems,” IEEE Trans. Autom. Control, vol. 65, no. 5, pp. 2001–2015, 2020

    2001

  7. [13]

    Switching Topology for Resilient Consensus using Wi-Fi Signals,

    T. Wheeler, E. Bharathi, and S. Gil, “Switching Topology for Resilient Consensus using Wi-Fi Signals,” in Proc. Int. Conf. Robot. Autom. , 2019, pp. 2018–2024

    2019

  8. [15]

    Can Competition Outperform Collaboration? The Role of Misbehaving Agents,

    L. Ballotta, G. Como, J. S. Shamma, and L. Schenato, “Can Competition Outperform Collaboration? The Role of Misbehaving Agents,” IEEE Trans. Autom. Control, vol. 69, no. 4, pp. 2308–2323, 2024. 15

    2024

  9. [17]

    The Role of Confidence for Trust-Based Resilient Consensus,

    L. Ballotta and M. Yemini, “The Role of Confidence for Trust-Based Resilient Consensus,” in Proc. American Control Conf. , 2024, pp. 2822– 2829

    2024

  10. [18]

    Beyond Robustness: A Taxonomy of Approaches towards Resilient Multi-Robot Systems,

    A. Prorok, M. Malencia, L. Carlone, G. S. Sukhatme, B. M. Sadler, and V . Kumar, “Beyond Robustness: A Taxonomy of Approaches towards Resilient Multi-Robot Systems,” 2021

    2021

  11. [19]

    Resilient Asymptotic Consensus in Robust Networks,

    H. J. LeBlanc, H. Zhang, X. Koutsoukos, and S. Sundaram, “Resilient Asymptotic Consensus in Robust Networks,” IEEE J. Sel. Areas Commun., vol. 31, no. 4, pp. 766–781, 2013

    2013

  12. [20]

    Resilient Distributed Averaging,

    S. M. Dibaji, M. Safi, and H. Ishii, “Resilient Distributed Averaging,” in Proc. American Control Conf. , 2019, pp. 96–101

    2019

  13. [21]

    Graph-theoretic approaches for analyzing the resilience of distributed control systems: A tutorial and survey,

    M. Pirani, A. Mitra, and S. Sundaram, “Graph-theoretic approaches for analyzing the resilience of distributed control systems: A tutorial and survey,” Automatica, vol. 157, p. 111264, 2023

    2023

  14. [22]

    Triangular Networks for Resilient Formations,

    D. Saldaña, A. Prorok, M. F. M. Campos, and V . Kumar, “Triangular Networks for Resilient Formations,” in Distrib. Auton. Robot. Syst.: 13th Int. Symp., 2018, pp. 147–159

    2018

  15. [23]

    Formations for Resilient Robot Teams,

    L. Guerrero-Bonilla, A. Prorok, and V . Kumar, “Formations for Resilient Robot Teams,” IEEE Robot. Autom. Lett. , vol. 2, no. 2, pp. 841–848, 2017

    2017

  16. [24]

    Design Guarantees for Resilient Robot Formations on Lattices,

    L. Guerrero-Bonilla, D. Saldaña, and V . Kumar, “Design Guarantees for Resilient Robot Formations on Lattices,” IEEE Robot. Autom. Lett. , vol. 4, no. 1, pp. 89–96, 2019

    2019

  17. [25]

    Multi-Robot Adversarial Resilience using Control Barrier Functions,

    M. Cavorsi, B. Capelli, L. Sabattini, and S. Gil, “Multi-Robot Adversarial Resilience using Control Barrier Functions,” in Proc. Robot. Sci. Syst. , vol. 18, 2022

    2022

  18. [26]

    Resilient Leader-Follower Consensus to Arbitrary Reference Values in Time-Varying Graphs,

    J. Usevitch and D. Panagou, “Resilient Leader-Follower Consensus to Arbitrary Reference Values in Time-Varying Graphs,”IEEE Trans. Autom. Control, vol. 65, no. 4, pp. 1755–1762, 2020

    2020

  19. [27]

    Resiliency in dynamic leader–follower multiagent systems,

    H. Rezaee, T. Parisini, and M. M. Polycarpou, “Resiliency in dynamic leader–follower multiagent systems,” Automatica, vol. 125, p. 109384, 2021

    2021

  20. [28]

    Dynamic Resilient Containment Control in Multirobot Systems,

    M. Santilli, M. Franceschelli, and A. Gasparri, “Dynamic Resilient Containment Control in Multirobot Systems,” IEEE Trans. Robot., vol. 38, no. 1, pp. 57–70, 2022

    2022

  21. [29]

    Resilient Real-Valued Consensus in Spite of Mobile Malicious Agents on Directed Graphs,

    Y . Wang, H. Ishii, F. Bonnet, and X. Défago, “Resilient Real-Valued Consensus in Spite of Mobile Malicious Agents on Directed Graphs,” IEEE Trans. Parallel Distrib. Syst. , vol. 33, no. 3, pp. 586–603, 2022

    2022

  22. [30]

    Resilient Consensus Through Event-Based Communication,

    Y . Wang and H. Ishii, “Resilient Consensus Through Event-Based Communication,” IEEE Trans. Control Netw. Syst. , vol. 7, no. 1, pp. 471–482, 2020

    2020

  23. [31]

    A Sample- Based Algorithm for Approximately Testing r-Robustness of a Digraph,

    Y . Yi, Y . Wang, X. He, S. Patterson, and K. H. Johansson, “A Sample- Based Algorithm for Approximately Testing r-Robustness of a Digraph,” in Proc. IEEE Conf. Decis. Contol , 2022, pp. 6478–6483

    2022

  24. [32]

    Finite-Time Consensus on the Median Value With Robustness Properties,

    M. Franceschelli, A. Giua, and A. Pisano, “Finite-Time Consensus on the Median Value With Robustness Properties,” IEEE Trans. Autom. Control, vol. 62, no. 4, pp. 1652–1667, 2017

    2017

  25. [33]

    Median-Based Resilient Consensus Over Time-Varying Random Networks,

    Y . Shang, “Median-Based Resilient Consensus Over Time-Varying Random Networks,” IEEE Trans. Circuits Syst. II , vol. 69, no. 3, pp. 1203–1207, 2022

    2022

  26. [34]

    Trust is the Cure to Distributed Consensus with Adversaries,

    J. S. Baras and X. Liu, “Trust is the Cure to Distributed Consensus with Adversaries,” in Proc. Mediterranean Conf. Control Autom. , Akko, Israel, 2019, pp. 195–202

    2019

  27. [35]

    Resilient Consensus Based on Evidence Theory and Weight Correction,

    V . Bonagura, C. Fioravanti, G. Oliva, and S. Panzieri, “Resilient Consensus Based on Evidence Theory and Weight Correction,” in Proc. American Control Conf. , 2023, pp. 393–398

    2023

  28. [36]

    Improving Network Con- nectivity and Robustness Using Trusted Nodes With Application to Resilient Consensus,

    W. Abbas, A. Laszka, and X. Koutsoukos, “Improving Network Con- nectivity and Robustness Using Trusted Nodes With Application to Resilient Consensus,” IEEE Trans. Control Netw. Syst. , vol. 5, no. 4, pp. 2036–2048, 2018

    2018

  29. [37]

    Resilient Consensus with Mobile Detectors Against Malicious Attacks,

    C. Zhao, J. He, and J. Chen, “Resilient Consensus with Mobile Detectors Against Malicious Attacks,” IEEE Trans. Signal Inf. Process. Netw. , vol. 4, no. 1, pp. 60–69, 2018

    2018

  30. [38]

    SecureArray: Improving wifi security with fine-grained physical-layer information,

    J. Xiong and K. Jamieson, “SecureArray: Improving wifi security with fine-grained physical-layer information,” in Proc. Annu. Int. Conf. Mobile Comput. Netw., ser. MobiCom ’13. New York, NY , USA: Association for Computing Machinery, 2013, pp. 441–452

    2013

  31. [39]

    Trust modeling in multi-robot patrolling,

    C. Pippin and H. Christensen, “Trust modeling in multi-robot patrolling,” in Proc. Int. Conf. Robot. Autom. , 2014, pp. 59–66

    2014

  32. [40]

    How Physicality Enables Trust: A New Era of Trust-Centered Cyberphysical Systems,

    S. Gil, M. Yemini, A. Chorti, A. Nedi ´c, H. V . Poor, and A. J. Goldsmith, “How Physicality Enables Trust: A New Era of Trust-Centered Cyberphysical Systems,” 2023

    2023

  33. [41]

    Trustworthy Distributed Average Consensus,

    C. N. Hadjicostis and A. D. Domínguez-García, “Trustworthy Distributed Average Consensus,” in Proc. IEEE Conf. Decis. Contol , 2022, pp. 7403– 7408

    2022

  34. [42]

    Resilience to Malicious Activity in Distributed Optimization for Cyberphysical Systems,

    M. Yemini, A. Nedi ´c, S. Gil, and A. J. Goldsmith, “Resilience to Malicious Activity in Distributed Optimization for Cyberphysical Systems,” in Proc. IEEE Conf. Decis. Contol , 2022, pp. 4185–4192

    2022

  35. [43]

    Resilient Distributed Optimization for Multi-Agent Cyberphysical Systems,

    M. Yemini, A. Nedi ´c, A. J. Goldsmith, and S. Gil, “Resilient Distributed Optimization for Multi-Agent Cyberphysical Systems,” IEEE Trans. Autom. Control, 2025

    2025

  36. [44]

    Crowd Vetting: Rejecting Adversaries via Collaboration With Application to Multirobot Flocking,

    F. Mallmann-Trenn, M. Cavorsi, and S. Gil, “Crowd Vetting: Rejecting Adversaries via Collaboration With Application to Multirobot Flocking,” IEEE Trans. Robot. , vol. 38, no. 1, pp. 5–24, 2022

    2022

  37. [45]

    Consensus and Cooper- ation in Networked Multi-Agent Systems,

    R. Olfati-Saber, J. A. Fax, and R. M. Murray, “Consensus and Cooper- ation in Networked Multi-Agent Systems,” Proc. IEEE, vol. 95, no. 1, pp. 215–233, 2007

    2007

  38. [46]

    Noise Resilient Distributed Average Consensus Over Directed Graphs,

    V . Khatana and M. V . Salapaka, “Noise Resilient Distributed Average Consensus Over Directed Graphs,” IEEE Trans. Signal Inf. Process. Netw., vol. 9, pp. 770–785, 2023

    2023

  39. [47]

    Competition- Based Resilience in Distributed Quadratic Optimization,

    L. Ballotta, G. Como, J. S. Shamma, and L. Schenato, “Competition- Based Resilience in Distributed Quadratic Optimization,” in Proc. IEEE Conf. Decis. Contol , 2022, pp. 6454–6459

    2022

  40. [48]

    Social influence and opinions,

    N. E. Friedkin and E. C. Johnsen, “Social influence and opinions,” J. Math. Sociol., vol. 15, no. 3-4, pp. 193–206, 1990

    1990

  41. [49]

    Opinion evolution in time-varying social influence networks with prejudiced agents,

    A. V . Proskurnikov, R. Tempo, M. Cao, and N. E. Friedkin, “Opinion evolution in time-varying social influence networks with prejudiced agents,” IFAC-PapersOnLine, vol. 50, no. 1, pp. 11 896–11 901, 2017

    2017

  42. [50]

    Convergence in Multiagent Coordination, Consensus, and Flocking,

    V . Blondel, J. Hendrickx, A. Olshevsky, and J. Tsitsiklis, “Convergence in Multiagent Coordination, Consensus, and Flocking,” in Proc. IEEE Conf. Decis. Contol , 2005, pp. 2996–3000

    2005

  43. [51]

    Distributed Average Consensus with Time-Varying Metropolis Weights,

    L. Xiao, S. Boyd, and S. Lall, “Distributed Average Consensus with Time-Varying Metropolis Weights,” Automatica, 2006

    2006

  44. [52]

    A Consensus-Based Approach for Platooning with Intervehicular Communications and Its Validation in Realistic Scenarios,

    S. Santini, A. Salvi, A. S. Valente, A. Pescapé, M. Segata, and R. Lo Cigno, “A Consensus-Based Approach for Platooning with Intervehicular Communications and Its Validation in Realistic Scenarios,” IEEE Trans. Veh. Technol., vol. 66, no. 3, pp. 1985–1999, 2017

    1985

  45. [53]

    Vehicular Platoon Communication: Architecture, Security Threats and Open Challenges,

    S. J. Taylor, F. Ahmad, H. N. Nguyen, and S. A. Shaikh, “Vehicular Platoon Communication: Architecture, Security Threats and Open Challenges,” Sensors, vol. 23, no. 1, p. 134, 2022

    2022

  46. [54]

    Conditional Convergence of Infinite Products,

    W. F. Trench, “Conditional Convergence of Infinite Products,” American Math. Monthly, vol. 106, no. 7, pp. 646–651, 1999

    1999

  47. [55]

    C. D. Meyer, Matrix Analysis and Applied Linear Algebra, Second Edition. Philadelphia, PA: Society for Industrial and Applied Mathematics, 2023

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.