A Lightweight Edge-Deployable ANN Model for Real-Time Energy Anomaly Detection in IoT-Driven Smart Grids
Abstract
The rapid expansion of the Internet of Things (IoT) in smart cities has necessitated efficient, real-time energy anomaly detection. However, complex hybrid deep learning models often exceed the computational capacity of Edge devices. This paper proposes a lightweight, 3-layer Artificial Neural Network (ANN) framework designed for Edge deployment. Using the LEAD (Large-scale Energy Anomaly Detection) dataset, we address class imbalance via the Synthetic Minority Over-sampling Technique (SMOTE). Our model achieves 98.4% accuracy, a macro F1-score of 0.93, and an AUC of 0.91. While these metrics are competitive with state-of-the-art hybrid models, our framework provides a significantly lower memory footprint and sub-millisecond inference latency, making it ideal for resource-constrained Edge environments.References
[1] V. Merlino and D. Allegra, “Energy-based approach for attack detection in IoT devices: A survey,” Internet of Things, vol. 27, p. 101306, Oct. 2024, doi: 10.1016/j.iot.2024.101306.
[2] Q. Vo, P. Ea, S. Benzouaoua, O. Salem, and A. Mehaoua, “Anomaly Detection in IoT Sensor Energy Consumption Using LSTM Neural Networks and Isolation Forest,” in 2024 7th Conference on Cloud and Internet of Things (CIoT), IEEE, Oct. 2024, pp. 1–8. doi: 10.1109/CIoT63799.2024.10756980.
[3] Y. Zhang, Y. Gao, and Z. Zhao, “Research on Operation and Anomaly Detection of Smart Power Grid Based on Information Technology Using
CNN+Bidirectional LSTM,” Informatica, vol. 49, no. 7, Feb. 2025, doi: 10.31449/inf.v49i7.7037.
[4] I. Priyadarshini, “Anomaly Detection of IoT Cyberattacks in Smart Cities Using Federated Learning and Split Learning,” Big Data and Cognitive Computing, vol. 8, no. 3, pp. 21–38, Feb. 2024, doi: 10.3390/bdcc8030021.
[5] Z. Hasani, S. Krrabaj, and M. Krasniqi, “Proposed Model for Real-Time Anomaly Detection in Big IoT Sensor Data for Smart City,” International Journal of Interactive Mobile Technologies (iJIM), vol. 18, no. 03, pp. 32–44, Feb. 2024, doi: 10.3991/ijim.v18i03.44467.
[6] T. Himdi and M. Ishaque, “Deep Learning-Enhanced anomaly detection for IoT security in smart cities,” ARPN Journal of Engineering and Applied Sciences, pp. 391–397, May 2024, doi: 10.59018/032456.
[7] H. R. O. Alghaithi, M. M. A. M. Alshehhi, and T. Murugan, “IoT Network Anomaly Detection Using Machine Learning and Deep Learning Techniques - Research Study,” in 2024 IEEE Students Conference on Engineering and Systems (SCES), IEEE, Jun. 2024, pp. 1–6. doi: 10.1109/SCES61914.2024.10652305.
[8] C. Ni, J. Wu, and H. Wang, “Energy-Aware Edge Computing Optimization for Real-Time Anomaly Detection in IoT Networks,” Applied and Computational Engineering, vol. 139, no. 1, pp. 42–53, Apr. 2025, doi: 10.54254/2755-2721/2025.22280.
[9] G. R. Kumar, A. D. Kulkarni, B. S. Kumar, N. Singh, V. Revathi, and T. Ch. A. Kumar, “Machine Learning Approaches for Anomaly Detection in IoT Networks,” in 2024 International Conference on Advances in Computing, Communication and Applied Informatics (ACCAI), IEEE, May 2024, pp. 1–5. doi: 10.1109/ACCAI61061.2024.10601954.
[10] P. Shobha Rani, M. Vajid Ahamed, K. S. Sai Chaithresh, S. Kundan Srinivas, and P. V. Vivek, “Utilizing Machine Learning Techniques for Detecting Anomalies in IoT Networks,” in 2024 5th International Conference on Recent Trends in Computer Science and Technology (ICRTCST), IEEE, Apr. 2024, pp. 105–110. doi: 10.1109/ICRTCST61793.2024.10578424.
[11] H. Bilakanti, S. Pasam, V. Palakollu, and S. Utukuru, “Anomaly detection in IoT environment using machine learning,” SECURITY AND PRIVACY, vol. 7, no. 3, May 2024, doi: 10.1002/spy2.366.
[12] M. Siddiqui, M. Asifuddola, M. Kalra, C. R. Krishna, and A. R. Khan, “Fog Enabled Anomaly Detection System for Sensors’ Anomaly in IoT Environment Using Machine Learning,” International Journal of System Assurance Engineering and Management, pp. 1–42, Mar. 2025, doi: 10.21203/rs.3.rs-5299588/v1.
[13] M. J. C. S. Reis, “AI-Driven Anomaly Detection for Securing IoT Devices in 5G-Enabled Smart Cities,” Electronics (Basel), vol. 14, no. 12, pp. 2492–2527, Jun. 2025, doi: 10.3390/electronics14122492.
[14] V. L. Chaitanya, G.Anju Sree, D. A. Thabusum, U.Sravani, G.Sneha, and K.Jyotshna, “Outlier Detection for IoT Frameworks using Machine Learning Techniques,” International Research Journal of Innovations in Engineering and Technology, vol. 09, no. Special Issue, pp. 180–184, 2025, doi: 10.47001/IRJIET/2025.INSPIRE30.
[15] S. H. Rafique, A. Abdallah, N. S. Musa, and T. Murugan, “Machine Learning and Deep Learning Techniques for Internet of Things Network Anomaly Detection—Current Research Trends,” Sensors, vol. 24, no. 6, pp. 1968–2000, Mar. 2024, doi: 10.3390/s24061968.
[16] P. Pustelnyk and Y. Levus, “Real-time anomaly detection in distributed IOT systems: a comprehensive review and comparative analysis,” Vìsnik Nacìonalʹnogo unìversitetu “Lʹvìvsʹka polìtehnìka”. Serìâ Ìnformacìjnì sistemi ta merežì, vol. 17, pp. 160–169, Jun. 2025, doi: 10.23939/sisn2025.17.160.
[17] P. Satish, Ch. Venu Yadav, V. Raj Kumar, B. Sai Krishna Reddy, and T. Vamshi Krishna, “EDGE-ENABLED MACHINE LEARNING FRAMEWORK FOR REALTIME ANOMALY DETECTION IN IOT NETWORK,” International Journal of Engineering Research and Science & Technology, vol. 21, no. 3 (1), pp. 1424–1431, Aug. 2025, doi: 10.62643/ijerst.v21.n3(1).pp1424-1431.
DOI:
https://doi.org/10.31449/inf.v50i1.12670Downloads
Published
How to Cite
Issue
Section
License
Authors retain copyright in their work. By submitting to and publishing with Informatica, authors grant the publisher (Slovene Society Informatika) the non-exclusive right to publish, reproduce, and distribute the article and to identify itself as the original publisher.
All articles are published under the Creative Commons Attribution license CC BY 3.0. Under this license, others may share and adapt the work for any purpose, provided appropriate credit is given and changes (if any) are indicated.
Authors may deposit and share the submitted version, accepted manuscript, and published version, provided the original publication in Informatica is properly cited.
