Anomaly Detection in Resilient Networks: On the Impact of Delayed and False Decisions

Laue F, Lübke M (2026)

Publication Type: Conference contribution

Publication year: 2026

Publisher: IEEE

City/Town: New York City

Pages Range: 1501-1507

Conference Proceedings Title: 2025 59th Asilomar Conference on Signals, Systems, and Computers

Event location: Pacific Grove US

DOI: 10.1109/IEEECONF67917.2025.11443750

Abstract

This paper studies a resilient communication network that monitors a key performance indicator (KPI) to detect anomalies and disruptions. Upon detection, the network automatically reconfigures to a recovery state, ensuring reduced but stable performance. To analyze the impact of the detection and recovery process on the network resilience, we propose two discrete-time Markov chains (DTMCs) that model the network dynamics under both normal and challenging conditions. The proposed DTMCs account for the effects of delayed and false decisions in the detection process and model the recovery capabilities of the considered network. Based on the statistical properties of the DTMCs, we derive a cost function to evaluate the long-term performance of the network, serving as a measure for resilience. In addition, we evaluate the performance of an exemplary network using a Gaussian signal model for the KPI and a jamming detector from the literature. Our results reveal that the proposed cost function is minimized at low probabilities of false alarm and relatively high probabilities of miss detection, indicating that false decisions can significantly affect the network performance. Furthermore, our evaluation shows that shorter measurement intervals for the KPI and faster recovery from disruptions improve the resilience of the considered network.

Authors with CRIS profile

Friedemann Laue Lehrstuhl für Intelligente Technische Elektronik und Systeme Maximilian Lübke Lehrstuhl für Intelligente Technische Elektronik und Systeme

How to cite

APA:

Laue, F., & Lübke, M. (2025). Anomaly Detection in Resilient Networks: On the Impact of Delayed and False Decisions. In 2025 59th Asilomar Conference on Signals, Systems, and Computers (pp. 1501-1507). Pacific Grove, US: New York City: IEEE.

MLA:

Laue, Friedemann, and Maximilian Lübke. "Anomaly Detection in Resilient Networks: On the Impact of Delayed and False Decisions." Proceedings of the 2025 59th Asilomar Conference on Signals, Systems, and Computers, Pacific Grove New York City: IEEE, 2025. 1501-1507.

BibTeX: Download