arxiv: 2605.13500 · v1 · submitted 2026-05-13 · 💻 cs.RO · cs.CR

Recognition: unknown

Uncertainty-Aware 3D Position Refinement for Multi-UAV Systems

Hosam Alamleh , Damir Pulatov

Authors on Pith no claims yet

Pith reviewed 2026-05-14 18:52 UTC · model grok-4.3

classification 💻 cs.RO cs.CR

keywords multi-UAV localization3D position refinementuncertainty-aware fusiontrust scoremalicious node detectiondecentralized systemsrange constraints

0 comments

The pith

Uncertainty-aware fusion of neighbor ranges and trust scores refines 3D UAV positions to lower localization errors.

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

This paper introduces a lightweight refinement layer that each UAV runs locally to improve its 3D position estimate. It combines the UAV's own uncertain prior with information from neighbors, weighted by reported covariances, ranging quality, and a trust score. The approach explicitly manages cases where local estimates are poor at startup or after loss of signal by relying more on trusted neighbors. It also detects and down-weights inconsistent or malicious neighbors using a range consistency check. The result is better performance in simulations when starting up or when some nodes are faulty, which matters for safe swarm navigation in real environments where GNSS can be unreliable.

Core claim

The paper claims that a decentralized position-refinement method, which performs uncertainty-aware neighborhood fusion by weighting priors according to covariance and neighbor constraints by link quality and a learned trust score, together with a local range-consistency check to exclude faulty neighbors, substantially reduces mean localization error during cold start and maintains lower error in the presence of malicious nodes.

What carries the argument

The uncertainty-aware neighborhood fusion mechanism that weights each contribution by covariance, ranging uncertainty, and trust score while applying a smoothed range-consistency check.

If this is right

Substantially reduces mean localization error during cold start with 10 UAVs.
Remains competitive with local estimators after they stabilize.
Maintains lower error as the fraction of malicious nodes increases compared to fusion without trust.

Where Pith is reading between the lines

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

Such a layer could be added to existing UAV navigation stacks to improve resilience without requiring new hardware.
Testing in real flights with actual range sensors would show whether the simulated gains hold under noise and timing delays.
Scaling the approach to hundreds of UAVs might require adjustments to communication overhead from sharing state summaries.

Load-bearing premise

The framework assumes that reliable inter-UAV range measurements are available and that the learned trust score together with the range-consistency check can identify and down-weight malicious neighbors without adding new errors.

What would settle it

A controlled simulation or flight test in which the mean 3D localization error with the proposed refinement exceeds the error from baseline fusion without the trust and consistency mechanisms would falsify the claimed improvement.

Figures

Figures reproduced from arXiv: 2605.13500 by Damir Pulatov, Hosam Alamleh.

**Figure 2.** Figure 2: Representative snapshot illustrating location refinement for a representative run (seed=570687052). The refined positi [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 4.** Figure 4: Robustness results: final-epoch mean 3D localization error of honest UAVs as a function of the malicious-node fraction. Trust-based mitigation consistently reduces error compared to no-trust fusion. all tested fractions. For instance, at 40% malicious nodes, the trust mechanism reduces the mean error from 6.99 m to 5.19 m . Additionally, the 10th–90th percentile bands remain tight even under high threat le… view at source ↗

**Figure 3.** Figure 3: Mean 3D localization error of honest UAVs over 30 epochs for 100 runs. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

read the original abstract

Reliable real-time 3D localization is essential for multi-UAV navigation, collision avoidance, and coordinated flight, yet onboard estimates can degrade under GNSS multipath, non-line-of-sight reception, vertical drift, and intentional interference. This paper presents a decentralized, lightweight 3D position-refinement layer that improves robustness by fusing each Unmanned Aerial Vehicle (UAV)'s local estimate with neighbor-shared state summaries and inter-UAV range or proximity constraints. The method performs uncertainty-aware neighborhood fusion by weighting each UAV's prior according to its reported covariance and weighting neighbor constraints according to link quality, ranging uncertainty, and a learned trust score. To support practical deployment, the framework explicitly handles cold start and temporary localization loss by inflating or substituting weak priors, allowing trusted neighborhood constraints to bootstrap and stabilize estimates until absolute sensing recovers. To mitigate the impact of faulty or malicious participants, each UAV applies a local range-consistency check, smoothed over time, to down-weight or exclude neighbors whose reported positions are incompatible with observed inter-UAV distances. Simulation experiments with 10 UAVs in a 3D volume show that the proposed refinement substantially reduces mean localization error during cold start, remains competitive after local estimators stabilize, and maintains lower error as the fraction of malicious nodes increases compared with fusion without trust. These results suggest that the approach can serve as a practical resilience layer for swarm operation in challenging 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.

Desk Editor's Note private letter to a colleague

The paper gives a workable decentralized fusion layer for UAV position refinement under GNSS trouble, but the simulation claims rest on thin evidence and untested assumptions about the trust and consistency checks.

read the letter

The core idea here is a lightweight, decentralized way to refine 3D UAV positions by blending each vehicle's local estimate with neighbor summaries and range measurements, while explicitly handling cold starts through prior inflation and using a learned trust score plus smoothed range-consistency checks to down-weight bad or malicious nodes. That combination is the main thing on offer, and it targets a real operational need in swarms where GNSS can drop or get jammed. The approach stays simple enough for onboard use and avoids heavy central coordination, which is a plus for practical deployment. The abstract shows simulation results with 10 UAVs claiming lower mean error during cold start and better resilience as malicious nodes increase, compared to plain fusion. That direction makes sense on paper. The soft spots are more noticeable. The results are described only at a high level with no error bars, statistical tests, or details on simulation conditions, parameter tuning, or how the learned trust score was trained. The central performance edge depends on the range-consistency check and trust mechanism correctly spotting faulty reports without letting consistent-but-biased data through, yet there's no sensitivity analysis or adversarial testing shown for coordinated attacks. The assumption that inter-UAV ranges are reliably available also sits in the background without much probing. This is the kind of work that could interest people building multi-UAV navigation stacks who need a resilience add-on rather than a full new localization system. It deserves a serious referee because the problem is relevant and the method is concrete enough to evaluate, even if the current evidence is preliminary. I'd send it for review with a request for expanded methods and results sections.

Referee Report

2 major / 2 minor

Summary. The manuscript presents a decentralized, lightweight 3D position-refinement layer for multi-UAV systems that fuses each UAV's local estimate with neighbor-shared state summaries and inter-UAV range or proximity constraints. The method performs uncertainty-aware neighborhood fusion by weighting priors according to reported covariance and neighbor constraints according to link quality, ranging uncertainty, and a learned trust score. It explicitly handles cold start and temporary localization loss by substituting weak priors with trusted neighborhood constraints and applies a local range-consistency check (smoothed over time) to down-weight or exclude incompatible neighbors. Simulation experiments with 10 UAVs in a 3D volume are reported to show substantially reduced mean localization error during cold start, competitiveness after local estimators stabilize, and maintained lower error as the fraction of malicious nodes increases compared with fusion without trust.

Significance. If the simulation results hold under more rigorous validation, the approach could provide a practical lightweight resilience layer for multi-UAV swarms operating in GNSS-challenged or interference-prone environments. It directly addresses cold-start bootstrapping and malicious-node mitigation through decentralized mechanisms that are relevant for real-time navigation and collision avoidance. The use of standard weighting combined with a learned trust score and consistency check offers a plausible engineering solution, though the current lack of statistical rigor and robustness testing limits the assessed impact.

major comments (2)

[Abstract (Simulation Experiments)] Abstract (Simulation Experiments): the abstract reports simulation outcomes with 10 UAVs but provides no quantitative error bars, statistical tests, or implementation details such as specific parameter choices for the learned trust score, simulation conditions, or number of runs; central performance claims rest on unspecified conditions and undermine assessment of the reported error reductions.
[Method (Range-Consistency Check and Trust Score)] Method (Range-Consistency Check and Trust Score): the claim that the range-consistency check plus learned trust score reliably down-weights malicious nodes (load-bearing for the increasing-malicious-fraction results) lacks sensitivity analysis or evaluation against coordinated attacks such as consistent-but-biased reports that satisfy observed ranges; if the check fails, bad data can propagate and increase rather than reduce error.

minor comments (2)

[Method] Clarify the exact mathematical formulation of the learned trust score, its update rule, and how it interacts with the range-consistency check in the fusion equations.
[Experiments] Simulation result figures should include error bars, specify the number of Monte Carlo runs, and report exact parameter values used for ranging uncertainty and trust-score learning.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our manuscript. We address each major comment point by point below, indicating where revisions will be made to improve clarity, rigor, and completeness.

read point-by-point responses

Referee: [Abstract (Simulation Experiments)] Abstract (Simulation Experiments): the abstract reports simulation outcomes with 10 UAVs but provides no quantitative error bars, statistical tests, or implementation details such as specific parameter choices for the learned trust score, simulation conditions, or number of runs; central performance claims rest on unspecified conditions and undermine assessment of the reported error reductions.

Authors: We agree that the abstract would be strengthened by including more quantitative support for the reported outcomes. In the revised manuscript we will update the abstract to report mean localization error reductions with standard deviation error bars across runs, specify that results are averaged over 50 Monte Carlo trials, and briefly note key simulation parameters (3D volume dimensions, UAV velocity bounds, and ranging noise model) along with the trust-score training regime (supervised learning on simulated trajectories with injected faults). Full algorithmic parameters and statistical test details will be expanded in the Experiments section to respect abstract length limits while making the central claims more readily assessable. revision: yes
Referee: [Method (Range-Consistency Check and Trust Score)] Method (Range-Consistency Check and Trust Score): the claim that the range-consistency check plus learned trust score reliably down-weights malicious nodes (load-bearing for the increasing-malicious-fraction results) lacks sensitivity analysis or evaluation against coordinated attacks such as consistent-but-biased reports that satisfy observed ranges; if the check fails, bad data can propagate and increase rather than reduce error.

Authors: The referee correctly identifies an important gap in the current evaluation. While the smoothed range-consistency check is intended to detect and down-weight neighbors whose reported states are incompatible with observed inter-UAV distances over time, the manuscript does not present sensitivity analysis against coordinated attacks in which malicious nodes issue mutually consistent but biased position reports that remain compatible with the measured ranges. We will add a new subsection in the Experiments section that evaluates the trust mechanism under such coordinated attack scenarios, reports the resulting localization error curves, and discusses residual failure modes together with possible additional safeguards. This addition will provide a more rigorous assessment of the method's robustness. revision: yes

Circularity Check

0 steps flagged

No circularity: method uses standard weighting and checks without self-referential derivations

full rationale

The paper describes a decentralized fusion approach relying on reported covariances, link quality, ranging uncertainty, a learned trust score, and a local range-consistency check. No equations, derivations, or parameter-fitting steps are shown in the provided text that reduce the claimed error reductions to quantities defined by the same data or inputs. The central claims rest on external measurements and standard consistency mechanisms rather than any self-definitional or fitted-input reductions, making the derivation self-contained against the described inputs.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on domain assumptions about range measurement availability and the effectiveness of learned trust scoring; no free parameters are explicitly quantified in the abstract, but trust-score learning and link-quality weights are implied tuning elements.

free parameters (2)

learned trust score
Trust score is described as learned; its training data and fitting procedure are unspecified.
link quality and ranging uncertainty weights
Weights are applied according to link quality and ranging uncertainty; exact functional form or calibration is not provided.

axioms (2)

domain assumption Inter-UAV range or proximity measurements are available and sufficiently accurate to serve as constraints
Invoked when weighting neighbor constraints and performing consistency checks.
domain assumption Local covariance estimates from onboard estimators are meaningful and can be used for weighting
Used to weight each UAV's prior during fusion.

pith-pipeline@v0.9.0 · 5551 in / 1213 out tokens · 84129 ms · 2026-05-14T18:52:44.668594+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages

[1]

Unmanned aerial systems for photogrammetry and remote sensing: A review,

I. Colomina and P . Molina, “Unmanned aerial systems for photogrammetry and remote sensing: A review,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 92, pp. 79–97, 2014

work page 2014
[2]

A survey of multi -agent formation control,

K.-K. Oh, M.-C. Park, and H. -S. Ahn, “A survey of multi -agent formation control,” Automatica, vol. 53, pp. 424–440, 2015

work page 2015
[3]

M. S. Grewal, A. P . Andrews, and C. G. Bartone, Global Navigation Satellite Systems, Inertial Navigation, and Integration , 4th ed. Wiley, 2020

work page 2020
[4]

Visual slam for unmanned aerial vehicles: Localization and perception,

L. Zhuang, X. Zhong, L. Xu, C. Tian, and W . Yu, “Visual slam for unmanned aerial vehicles: Localization and perception,” Sensors, vol. 24, no. 10, p. 2980, 2024

work page 2024
[5]

Barometric altitude measurement fault diagnosis for the improvement of quadcopter altitude control,

N. Xuan-Mung and S. K. Hong, “Barometric altitude measurement fault diagnosis for the improvement of quadcopter altitude control,” in 2019 19th International Conference on Control, Automation and Systems (ICCAS), Jeju, Korea, 2019, pp. 1359–1364

work page 2019
[6]

On GPS spoofing of aerial platforms: A review of threats, challenges, methodologies, and future research directions,

S. Z. Khan, M. Mohsin, and W . Iqbal, “On GPS spoofing of aerial platforms: A review of threats, challenges, methodologies, and future research directions,” PeerJ Computer Science, vol. 7, p. e507, 2021

work page 2021
[7]

Advances in uav detection: Integrating multisensor systems and ai for enhanced accuracy and efficiency,

V. Semenyuk, I. Kurmashev, A. Lupidi, D. Alyoshin, L. Kurmasheva, and A. Cantelli-Forti, “Advances in uav detection: Integrating multisensor systems and ai for enhanced accuracy and efficiency,” International Journal of Critical Infrastructure Protection, vol. 49, p. 100744, 2025

work page 2025
[8]

Cooperative UWB-based localization for outdoors positioning and navigation of UAVs aided by ground robots,

X. Yu, Q. Li, J. P . Queralta, J. Heikkonen, and T . Westerlund, “Cooperative UWB-based localization for outdoors positioning and navigation of UAVs aided by ground robots,” in 2021 IEEE International Conference on Autonomous Systems (ICAS) . IEEE, 2021, pp. 1 –5, also available as arXiv:2104.00302

work page arXiv 2021
[9]

Cooperative localization in wireless networks,

H. Wymeersch, J. Lien, and M. Z. Win, “Cooperative localization in wireless networks,” Proceedings of the IEEE, vol. 97, no. 2, pp. 427– 450, 2009

work page 2009
[10]

Secure positioning in wireless networks,

S. Capkun and J. Hubaux, “Secure positioning in wireless networks,” IEEE Journal on Selected Areas in Communications , vol. 24, no. 2, pp. 221 – 232, 2006

work page 2006
[11]

The development of distributed cooperative localization algorithms for multi -uav systems in the past decade,

Z. Zhang, N. Li, G. Yan, and W . Li, “The development of distributed cooperative localization algorithms for multi -uav systems in the past decade,” Measurement, vol. 256, p. 118040, 2025

work page 2025
[12]

Uwb -based localization for multi -uav systems and collaborative heterogeneous multi -robot systems,

S. Wang, C. Mart´ınez Almansa, J. Pena Queralta, Z. Zou, and T . Wester-˜ lund, “Uwb -based localization for multi -uav systems and collaborative heterogeneous multi -robot systems,” Procedia Computer Science , vol. 175, pp. 357–364, 2020

work page 2020
[13]

An ultra -widebandbased multi-uav localization system in gps -denied environments,

T . M. Nguyen, A. H. Zaini, K. Guo, and L. Xie, “An ultra -widebandbased multi-uav localization system in gps -denied environments,” in International Micro Air Vehicle Conference and Competition (IMAV) , 2016

work page 2016
[14]

Ultra -wideband based cooperative relative localization algorithm and experiments for multiple unmanned aerial vehicles in gps denied environments,

K. Guo, Z. Qiu, W . Meng, L. Xie, and R. Teo, “Ultra -wideband based cooperative relative localization algorithm and experiments for multiple unmanned aerial vehicles in gps denied environments,” International Journal of Micro Air Vehicles, vol. 9, no. 3, pp. 169–186, 2017

work page 2017
[15]

Ultra -wideband and odometry -based cooperative relative localization with application to multi-uav formation control,

K. Guo, X. Li, and L. Xie, “Ultra -wideband and odometry -based cooperative relative localization with application to multi-uav formation control,” IEEE Transactions on Cybernetics, vol. 50, no. 6, pp. 2590– 2603, 2020

work page 2020
[16]

Vio -uwb-based collaborative localization and dense scene reconstruction within heterogeneous multi-robot systems,

J. P . Queralta, Q. Li, F. Schiano, and T . Westerlund, “Vio -uwb-based collaborative localization and dense scene reconstruction within heterogeneous multi-robot systems,” in 2022 International Conference on Advanced Robotics and Mechatronics (ICARM). IEEE, 2022, pp. 87– 94

work page 2022
[17]

Cooperative localization against gps signal loss in multiple uavs flight,

Y . Qu and Y . Zhang, “Cooperative localization against gps signal loss in multiple uavs flight,” Journal of Systems Engineering and Electronics, vol. 22, no. 1, pp. 103–112, Feb. 2011

work page 2011
[18]

Gnss spoofing and detection,

M. L. Psiaki and T . E. Humphreys, “Gnss spoofing and detection,” Proceedings of the IEEE, vol. 104, no. 6, pp. 1258–1270, 2016

work page 2016
[19]

Hirloc: high -resolution robust localization for wireless sensor networks,

L. Lazos and R. Poovendran, “Hirloc: high -resolution robust localization for wireless sensor networks,” IEEE Journal on Selected Areas in Communications, vol. 24, no. 2, pp. 233–246, 2006

work page 2006
[20]

Resilient distributed estimation: Sensor attacks,

Y . Chen, S. Kar, and J. M. F. Moura, “Resilient distributed estimation: Sensor attacks,” IEEE Transactions on Automatic Control , vol. 64, no. 9, Nov. 2018

work page 2018
[21]

Resilient distributed state estimation with mobile agents: overcoming byzantine adversaries, communication losses, and intermittent measurements,

A. Mitra, J. A. Richards, S. Bagchi, and S. Sundaram, “Resilient distributed state estimation with mobile agents: overcoming byzantine adversaries, communication losses, and intermittent measurements,” Autonomous Robots, vol. 43, no. 3, pp. 743–768, 2019

work page 2019
[22]

Uncertainty -aware 3d position refinement for multi -uav systems,

H. Alamleh and D. Pulatov, “Uncertainty -aware 3d position refinement for multi -uav systems,” https://github.com/hosam37r/ Uncertainty - Aware-3D-Position-Refinement-for-Multi-UAV-Systems, 2026, gitHub repository, accessed 2026-04-15

work page 2026