A Study on Distributed Learning for Attack Detection in Wireless Sensor Networks

Wireless Sensor Networks (WSNs) have become a cornerstone in the implementation of Structural Health Monitoring (SHM) systems due to their versatility, energy efficiency, and adaptability. Their ability to facilitate low-powered communication makes them ideal for monitoring the structural integrity...

Full description

Saved in:
Bibliographic Details
Published in:Prognostics and System Health Management Conference pp. 1 - 8
Main Authors: Savage, Eric Noah, Popillo, Cameron, Zhou, Ruolin
Format: Conference Proceeding
Language:English
Published: IEEE 09.06.2025
Subjects:
Computational modeling
Computer hacking
Fault tolerance
Heterogeneous networks
Intrusion detection
Monitoring
Predictive models
Prevention and mitigation
Prognostics and health management
Wireless sensor networks
ISSN:2166-5656
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Wireless Sensor Networks (WSNs) have become a cornerstone in the implementation of Structural Health Monitoring (SHM) systems due to their versatility, energy efficiency, and adaptability. Their ability to facilitate low-powered communication makes them ideal for monitoring the structural integrity of environments where traditional wired connections are impractical. This paper explores the integration of a computer cluster with a WSN to enhance the functionality of an SHM system. Specifically, we investigate the feasibility of using a cluster of four Raspberry Pi computers to assess the feasibility of using distributed models in PyTorch and TensorFlow, assessing their applicability in heterogeneous network environments. Additionally, we examine the differences between machine learning-based and non-machine learning-based security approaches for WSNs. We analyze existing attack mitigation strategies, including traditional rule-based intrusion detection systems and cryptographic techniques, alongside ML-driven anomaly detection, classification, and threat prediction models. By evaluating how a computer cluster can enhance these security mechanisms, we aim to provide insights into the effectiveness of ML and non-ML-based techniques in improving attack detection, prevention, and overall system resilience in SHM deployments.
ISSN:2166-5656
DOI:10.1109/ICPHM65385.2025.11062064