FL-EDGEWAD: A FEDERATED LEARNING-BASED EDGE FRAMEWORK FOR REAL-TIME AND PRIVACY-PRESERVING WASTEWATER ANOMALY DETECTION
Keywords
DOI:
https://doi.org/10.34238/tnu-jst.15224Abstract
Rapid urbanization and industrial activities have increased the need for continuous and reliable wastewater quality monitoring to prevent environmental risks and ensure regulatory compliance. However, conventional centralized IoT-based systems often suffer from high latency, significant communication overhead, and potential data privacy concerns due to centralized data collection. In addition, many existing approaches lack effective mechanisms for real-time anomaly detection across distributed monitoring sites while preserving local data confidentiality. To address these challenges, this paper proposes FL-EdgeWAD, a federated learning-based edge intelligence framework for collaborative anomaly detection. The framework integrates a lightweight Long Short-Term Memory Autoencoder (LSTM-AE) deployed on resource-constrained edge devices with the Federated Averaging (FedAvg) algorithm to enable privacy-preserving global model training. Experiments were conducted on a simulated multivariate wastewater dataset comprising 50,000 samples collected from five distributed IoT nodes, monitoring four water quality parameters: pH, dissolved oxygen, total dissolved solids, and temperature. Anomalies accounting for approximately 5% of total samples were injected to simulate real-world pollution events. The results show that FL-EdgeWAD achieves an accuracy of 92.3% and an F1-score of 91.5%, outperforming a local LSTM-AE baseline by 7.1% while converging within 25 communication rounds. These findings demonstrate that the proposed framework provides an effective balance between scalability, latency, and data privacy, making it suitable for real-world intelligent environmental monitoring applications.Downloads
References
B. Ngwenya et al., "Monitoring ambient water quality using machine learning: A review of AWQMS and SDG 6.3.2," Science of The Total Environment, vol. 857, 2023, Art. no. 159502, doi: 10.1016/j.scitotenv.2022.159505.
S. Yang, K. Behzadian, C. Coleman, T. G. Holloway, and L. C. Campos, "Application of AI-based techniques for anomaly management in wastewater treatment plants: A review," Journal of Environmental Management, vol. 392, 2025, Art. no. 126886, doi: 10.1016/j.jenvman.2025.126886.
World Health Organization (WHO), Guidelines for Drinking-water Quality, 4th ed., Geneva, Switzerland, 2017.
E. El-Shafeiy et al., "Real-Time Anomaly Detection for Water Quality Sensor Data: A Systematic Literature Review," Sensors, vol. 23, no. 21, 2023, Art. no. 8847, doi: 10.3390/s23218847.
L. Da Xu, W. He, and S. Li, "Internet of Things in Industries: A Survey," IEEE Transactions on Industrial Informatics, vol. 10, no. 4, pp. 2233-2243, 2014, doi: 10.1109/TII.2014.2300753.
W. Shi et al., "Edge Computing: Vision and Challenges," IEEE Internet of Things Journal, vol. 3, no. 5, pp. 637-646, 2016, doi: 10.1109/JIOT.2016.2579198.
T. Li, A. K. Sahu, A. Talwalkar, and V. Smith, "Federated Learning: Challenges, Methods, and Future Directions," IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50-60, 2020, doi: 10.1109/MSP.2020.2975749.
P. Kairouz et al., "Advances and Open Problems in Federated Learning," Foundations and Trends in Machine Learning, vol. 14, no. 1-2, 2021, doi: 10.1561/2200000083.
M. N. A. Ramadan et al., "SecureIoT-FL: A Federated Learning Framework for Privacy-Preserving Real-Time Environmental Monitoring in Industrial IoT Applications," Alexandria Engineering Journal, vol. 114, pp. 681-701, 2025, doi: 10.1016/j.aej.2024.11.069.
H. B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. Y. Arcas, "Communication-Efficient Learning of Deep Networks from Decentralized Data," in Proc. AISTATS, 2017, pp. 1273-1282.
A. Alwabli et al., "Federated Learning for Privacy-Preserving Air Quality Forecasting using IoT Sensors," Engineering, Technology & Applied Science Research, vol. 14, no. 2, pp. 13001-13008, 2024, doi: 10.48084/etasr.6833.
Y. Wei et al., "LSTM-Autoencoder-Based Anomaly Detection for Indoor Air Quality Time-Series Data," IEEE Sensors Journal, vol. 23, no. 4, pp. 4025-4036, 2023, doi: 10.1109/JSEN.2022.3230361.
M. N. Alatawi, "Edge computing and federated learning for privacy-preserving IoT analytics," Journal on Wireless Communications and Networking, vol. 2025, 2025, Art. no. 24545, doi: 10.1186/s13638-025-02545.
P. Warden and D. Situnayake, TinyML: Machine Learning with TensorFlow Lite on Arduino and Ultra-Low-Power Microcontrollers, O'Reilly Media, 2019.
