FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection

As distributed sensing technologies evolve, the collection of time series data is becoming increasingly decentralized, which introduces serious challenges for both model training and data privacy protection. In response to this trend, federated time series anomaly detection enables collaborative ana...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiuxian Zhang, Hongwei Zhao, Weishan Zhang, Shaohua Cao, Haoyun Sun, Baoyu Zhang
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/13/4014
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849704271041265664
author Xiuxian Zhang
Hongwei Zhao
Weishan Zhang
Shaohua Cao
Haoyun Sun
Baoyu Zhang
author_facet Xiuxian Zhang
Hongwei Zhao
Weishan Zhang
Shaohua Cao
Haoyun Sun
Baoyu Zhang
author_sort Xiuxian Zhang
collection DOAJ
description As distributed sensing technologies evolve, the collection of time series data is becoming increasingly decentralized, which introduces serious challenges for both model training and data privacy protection. In response to this trend, federated time series anomaly detection enables collaborative analysis across distributed sensing nodes without exposing raw data. However, federated anomaly detection experiences issues with unstable training and poor generalization due to client heterogeneity and the limited expressiveness of single-path detection methods. To address these challenges, this study proposes FedSW-TSAD, a federated time series anomaly detection method based on the Sobolev–Wasserstein GAN (SWGAN). It leverages the Sobolev–Wasserstein constraint to stabilize adversarial training and combines discriminative signals from both reconstruction and prediction modules, thereby improving robustness against diverse anomalies. In addition, FedSW-TSAD adopts a differential privacy mechanism with L2-norm-constrained noise injection, ensuring privacy in model updates under the federated setting. The experimental results determined using four real-world sensor datasets demonstrate that FedSW-TSAD outperforms existing methods by an average of 14.37% in the F1-score while also enhancing gradient privacy under the differential privacy mechanism. This highlights the practical value of FedSW-TSAD for privacy-preserving anomaly detection in sensor-based monitoring systems such as industrial IoT, remote diagnostics, and predictive maintenance.
format Article
id doaj-art-0b14ac57e793456e9f5cd3e1d8f65bb8
institution DOAJ
issn 1424-8220
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-0b14ac57e793456e9f5cd3e1d8f65bb82025-08-20T03:16:50ZengMDPI AGSensors1424-82202025-06-012513401410.3390/s25134014FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly DetectionXiuxian Zhang0Hongwei Zhao1Weishan Zhang2Shaohua Cao3Haoyun Sun4Baoyu Zhang5Qingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao 266580, ChinaQingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao 266580, ChinaQingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao 266580, ChinaQingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao 266580, ChinaQingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao 266580, ChinaQingdao Institute of Software, College of Computer Science and Technology, China University of Petroleum (East China), Qingdao 266580, ChinaAs distributed sensing technologies evolve, the collection of time series data is becoming increasingly decentralized, which introduces serious challenges for both model training and data privacy protection. In response to this trend, federated time series anomaly detection enables collaborative analysis across distributed sensing nodes without exposing raw data. However, federated anomaly detection experiences issues with unstable training and poor generalization due to client heterogeneity and the limited expressiveness of single-path detection methods. To address these challenges, this study proposes FedSW-TSAD, a federated time series anomaly detection method based on the Sobolev–Wasserstein GAN (SWGAN). It leverages the Sobolev–Wasserstein constraint to stabilize adversarial training and combines discriminative signals from both reconstruction and prediction modules, thereby improving robustness against diverse anomalies. In addition, FedSW-TSAD adopts a differential privacy mechanism with L2-norm-constrained noise injection, ensuring privacy in model updates under the federated setting. The experimental results determined using four real-world sensor datasets demonstrate that FedSW-TSAD outperforms existing methods by an average of 14.37% in the F1-score while also enhancing gradient privacy under the differential privacy mechanism. This highlights the practical value of FedSW-TSAD for privacy-preserving anomaly detection in sensor-based monitoring systems such as industrial IoT, remote diagnostics, and predictive maintenance.https://www.mdpi.com/1424-8220/25/13/4014anomaly detectionprivacy protectionfederated learning
spellingShingle Xiuxian Zhang
Hongwei Zhao
Weishan Zhang
Shaohua Cao
Haoyun Sun
Baoyu Zhang
FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection
Sensors
anomaly detection
privacy protection
federated learning
title FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection
title_full FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection
title_fullStr FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection
title_full_unstemmed FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection
title_short FedSW-TSAD: SWGAN-Based Federated Time Series Anomaly Detection
title_sort fedsw tsad swgan based federated time series anomaly detection
topic anomaly detection
privacy protection
federated learning
url https://www.mdpi.com/1424-8220/25/13/4014
work_keys_str_mv AT xiuxianzhang fedswtsadswganbasedfederatedtimeseriesanomalydetection
AT hongweizhao fedswtsadswganbasedfederatedtimeseriesanomalydetection
AT weishanzhang fedswtsadswganbasedfederatedtimeseriesanomalydetection
AT shaohuacao fedswtsadswganbasedfederatedtimeseriesanomalydetection
AT haoyunsun fedswtsadswganbasedfederatedtimeseriesanomalydetection
AT baoyuzhang fedswtsadswganbasedfederatedtimeseriesanomalydetection