Alajmi, Abdullah, Petri, Ioan ORCID: https://orcid.org/0000-0002-1625-8247, Amin, Amin ORCID: https://orcid.org/0000-0002-6891-5640, Fadil, Fodil and Yacine, Rezgui ORCID: https://orcid.org/0000-0002-5711-8400
2026.
Distributed federated learning for secured urban ecosystems: A conceptual implementation for sports facilities.
Presented at: ECPPM 2026 – European Conference on Product and Process Modelling,
Cardiff, UK,
9-11 September 2026.
ECPPM 2026 – European Conference on Product and Process Modelling.
Zenodo,
10.5281/zenodo.22679960
|
Preview |
PDF
- Published Version
Available under License Creative Commons Attribution. Download (4MB) | Preview |
Abstract
The rapid deployment of Internet of Things (IoT) devices in smart urban ecosystems increases cybersecurity and privacy risks for connected infrastructure. This study proposes a distributed federated learning framework for IoT attack classification, using stadia and sports facilities as a conceptual implementation setting for mapping building subsystems to local federated clients. Logistic Regression (LR) and Long Short-Term Memory (LSTM) models are evaluated under centralised and federated settings. The models are assessed under Gaussian, Poisson, and Negative Binomial security-event distributions and benchmarked using cyber-attack classes. Results show that federated learning supports collaborative training and preserves data privacy while maintaining competitive classification performance. Federated LSTM generally outperforms LR-based models, especially for temporal or sequential attack patterns. The findings support federated learning as a scalable and privacy-preserving approach for securing urban IoT ecosystems.
| Item Type: | Conference or Workshop Item - published (Paper) |
|---|---|
| Date Type: | Publication |
| Status: | Published |
| Schools: | Schools > Engineering |
| Publisher: | Zenodo |
| Date of First Compliant Deposit: | 17 November 2026 |
| Last Modified: | 19 Nov 2026 14:59 |
| URI: | https://orca.cardiff.ac.uk/id/eprint/191002 |
Actions (repository staff only)
![]() |
Edit Item |





Dimensions
Dimensions