DYNAMIC TRUST ASSESSMENT METHOD USING MODIFIED HYSTERESIS IN UAV SWARMS

DOI: 10.31673/2412-4338.2026.019021

Abstract

The article investigates the problem of stability and security in the functioning of mobile complexes, particularly unmanned aerial vehicle (UAV) swarms, under conditions of data manipulation. Decentralized networks are vulnerable to On-Off attacks, during which a malicious node intentionally fluctuates its trust rating around the disconnection threshold. This causes a destabilizing effect of network state “flapping”, leading to an avalanche-like increase in traffic. An analysis of publications shows that existing solutions concentrate on global algorithms for calculating threat levels, leaving the issue of trust stabilization at the local (node) level unresolved. The objective of the work is to develop a mathematical model that serves as the basis for a dynamic trust assessment method: a mechanism for continuous updating of the trust coefficient and constant regulation of the nodes' effective weight using modified hysteresis. The main material describes the developed local stabilization module, which acts as a filter for external adaptive thresholds. The scientific novelty of the approach lies in the original application of the hysteresis mechanism for dynamic trust assessment tasks, where it is adapted for conditions in which a variable security threshold is fed to its input. The mathematical apparatus for trust management is based on exponential smoothing (EWMA) and hysteresis. The deviation duration counter and a binary quarantine indicator, which isolates the node from decision-making, are introduced. The proposed continuous piecewise-linear trust degradation function provides asymmetric penalization of the attacker, neutralizing attempts at rapid rating recovery. To verify the model, an isolated simulation of the node's behavior under an On-Off attack was conducted. The results prove the ability of modified hysteresis to independently block the violator's manipulations, eliminating the “flapping” effect. The conclusions confirm that simple trust fluidity is vulnerable without a stabilization apparatus. The proposed model successfully transforms the unstable behavior of agents into a secure, continuously managed effective weight suitable for calculating a secure consensus in UAV swarms. Prospects for future research involve optimizing the real-time interaction of the created local model with global neural network models.

Keywords: distributed systems, UAVs, dynamic trust, modified hysteresis, cybersecurity, On-Off attacks, continuous trust update

References

1. Adaptive multi-granularity trust management scheme for UAV visual sensor security under adversarial attacks / H. Li et al. Computers & security. 2024. P. 104108. URL: https://doi.org/10.1016/j.cose.2024.104108 (date of access: 23.03.2026).
2. A novel distributed situation awareness consensus approach for UAV swarm systems / X. Hai et al. IEEE transactions on intelligent transportation systems. 2023. P. 1–12. URL: https://doi.org/10.1109/tits.2023.3300871 (date of access: 23.03.2026).
3. Continuous authentication for UAV delivery systems under zero-trust security framework / C. Dong et al. 2022 IEEE international conference on edge computing and communications (EDGE), Barcelona, Spain, 10–16 July 2022. 2022. URL: https://doi.org/10.1109/edge55608.2022.00027 (date of access: 23.03.2026).
4. Deep neural network and zero trust principles for intelligent drone management / S. Q. Abbas et al. 2024 IEEE 9th international conference on engineering technologies and applied sciences (ICETAS), Bahrain, Bahrain, 20–22 November 2024. 2024. P. 1–6. URL: https://doi.org/10.1109/icetas62372.2024.11120042 (date of access: 23.03.2026).
5. Kannan S., Venkataraman R., Ramachandran G. S. On-off attack detection in trust model using intra-daily variability for the IoT. Bulletin of electrical engineering and informatics. 2023. Vol. 12, no. 6. P. 3880–3888. URL: https://doi.org/10.11591/eei.v12i6.6110 (date of access: 23.03.2026).
6. Liao Z., Zhang L., Dong Z. UAV swarm exploration with byzantine fault tolerance. 2021 china automation congress (CAC), Beijing, China, 22–24 October 2021. 2021. URL: https://doi.org/10.1109/cac53003.2021.9727874 (date of access: 23.03.2026).
7. Reputation-Enhanced practical byzantine fault tolerance algorithm for node capture attacks on UAV networks / P. Onukak et al. Discover internet of things. 2025. Vol. 5, no. 1. URL: https://doi.org/10.1007/s43926-025-00164-y (date of access: 23.03.2026).
8. Saffre F., Hildmann H., Anttonen A. Force-Based Self-Organizing MANET/FANET with a UAV Swarm. Future internet. 2023. Vol. 15, no. 9. P. 315. URL: https://doi.org/10.3390/fi15090315 (date of access: 23.03.2026).
9. Sharma J., Mehra P. S. A survey of security challenges and existing prevention methods in FANET. Intelligent data analytics, iot, and blockchain. Boca Raton, 2023. P. 252–262. URL: https://doi.org/10.1201/9781003371380-24 (date of access: 23.03.2026).
10. Singh R. Simulation and evaluation of a hybrid trust–cryptographic protocol for UAV swarm communications. Simulation modelling practice and theory. 2025. P. 103230. URL: https://doi.org/10.1016/j.simpat.2025.103230 (date of access: 23.03.2026).
11. The essential role of cybersecurity in UAV swarm operations / F. M. Khan et al. Unmanned aerial vehicles swarm for protecting smart cities. Berkeley, CA, 2025. P. 403–441. URL: https://doi.org/10.1007/979-8-8688-1047-3_11 (date of access: 23.03.2026).
12. Trust based flying ad-hoc network: a survey / J. Kundu et al. IEEE access. 2024. P. 1. URL: https://doi.org/10.1109/access.2024.3419904 (date of access: 23.03.2026).
13. Trust prediction based on grey exponential smoothing method in VANETs / S. Zhang et al. 12th EAI international conference on mobile multimedia communications, mobimedia 2019, 29th - 30th jun 2019, weihai, china, Weihai, People's Republic of China, 29–30 June 2019. 2019. URL: https://doi.org/10.4108/eai.29-6-2019.2282065 (date of access: 23.03.2026).
14. Xie G., Zhou X., Gao J. Adaptive trust threshold model based on reinforcement learning in cooperative spectrum sensing. Sensors. 2023. Vol. 23, no. 10. P. 4751. URL: https://doi.org/10.3390/s23104751 (date of access: 23.03.2026).
15. Zhou S., Zhang G., Meng X. LocTrust: A local and global consensus-combined trust model in MANETs. Peer-to-Peer networking and applications. 2021. Vol. 15, no. 1. P. 355–368. URL: https://doi.org/10.1007/s12083-021-01250-y (date of access: 23.03.2026).