RAVA: A Human-in-the-Loop Hybrid Platform for Explainable Anomaly Detection and Diagnosis in LTE RANs
Main Article Content
Keywords
LTE RAN Management, Intelligent Monitoring Platform, Human-in-the-Loop, Automated Root Cause Analysis, Real-Time Visualisation
Abstract
Today, Deep Learning (DL) has significantly improved anomaly detection accuracy in LTE Radio Access Networks (RANs). However, a gap remains between theoretical performance and real-world deployment. Traditional monitoring tools lack the interactivity and interpretability necessary for quick troubleshooting and resolution. In this paper, we present RAVA (Real-Time Anomaly Visualisation and Analysis), a comprehensive web platform that operationalises AI-driven detection. RAVA proposes a modular architecture that separates heavy inference tasks from visualisation, utilising WebSockets for real-time responsiveness. Our system includes a diagnostic engine that converts SHapley Additive exPlanations (SHAP)-based feature importance into actionable root causes, such as congestion and interference, through a deterministic, rule-based module. Additionally, a Human-in-the-Loop (HITL) feedback mechanism allows engineers to validate detections and actively improve the ground truth repository for ongoing system enhancement. The platform is tested with a real-world dataset of 1,650 eNodeBs. While relying on the underlying engine, which achieves a 93.89% F1-score, this study demonstrates its operational usefulness through a detailed case study of a site failure. Results indicate that RAVA effectively bridges the gap between black-box AI and engineering workflows. This platform reduces time-to-insight and supports closed-loop network management.
References
Basaran, O. T., & Dressler, F. (2025). XAInomaly: Explainable and interpretable Deep Contractive Autoencoder for O-RAN traffic anomaly detection. Computer Networks, 261, 111145. https://doi.org/10.1016/j.comnet.2025.111145
Bogucka, H., Hoffmann, M., Kryszkiewicz, P., & Kułacz, Ł. (2025). An Open-RAN Testbed for Detecting and Mitigating Radio-Access Anomalies. IEEE Communications Magazine, 63(11), 122–127. https://doi.org/10.1109/MCOM.003.2400513
Bordeau-Aubert, K., Whatley, J., Nadeau, S., Glatard, T., & Jaumard, B. (2023). Classification of anomalies in telecommunication network KPI time series. arXiv preprint, arXiv:2308.16279. https://doi.org/10.48550/arXiv.2308.16279
Dayaratne, T., Pham, N. D., Vo, V., Lai, S., Abuadbba, S., Suzuki, H., Yuan, X., & Rudolph, C. (2025). Robust Anomaly Detection in O-RAN: Leveraging LLMs against Data Manipulation Attacks. arXiv preprint, arXiv:2508.08029. https://doi.org/10.48550/arXiv.2508.08029
Kahoul, A. E. M., Bouraoui, Z., Zarour, K., & Boumezbeur, I. (2025). Hybrid Deep Learning Ensemble with Dynamic Fusion for Reliable Anomaly Detection in Operational LTE RANs. Journal of Telecommunications and the Digital Economy, 13(4), 105–126. https://doi.org/10.18080/jtde.v13n4.1363
Lavanya, A., & K. Sekar, K. (2023). Traditional methods and machine learning for anomaly detection in self-organizing networks. International Journal of Scientific Research in Science, Engineering and Technology, 10(6), 352–360. https://doi.org/10.32628/IJSRSET2310662
Oğuz, H. T., & Kalaycıoğlu, A. (2022). Anomaly detection in multi-tiered cellular networks using LSTM and 1D CNN. EURASIP Journal on Wireless Communications and Networking, 2022(1), 101. https://doi.org/10.1186/s13638-022-02183-7
Safizadeh, A. (2024). Automated anomaly detection in telecom networks: Applying machine learning techniques to detect anomalies in multivariate time series (Unpublished master’s thesis). KTH EECS.
Sun, C., Pawar, U., Khoja, M., Foukas, X., Marina, M. K., & Radunovic, B. (2024). Spotlight: Accurate, explainable and efficient anomaly detection for open RAN. In Proceedings of the 30th ACM MobiCom (pp. 923–937). Association for Computing Machinery. https://doi.org/10.1145/3636534.3649380
Tanhatalab, M. R., Yousefi, H., Hosseini, H. M., Mofarah Bonab, M., Fakharian, V., & Abarghouei, H. (2019). Deep RAN: A scalable data-driven platform to detect anomalies in live cellular networks using recurrent convolutional neural networks. arXiv e-prints. https://doi.org/10.48550/arXiv.1911.04472
Tayeh, T., Aburakhia, S., Myers, R., & Shami, A. (2022). An attention-based ConvLSTM autoencoder with dynamic thresholding for unsupervised anomaly detection in multivariate time series. arXiv preprint, arXiv:2201.09172. https://doi.org/10.3390/make4020015
Zou, C., Yuan, A., & Hu, J. (2024). BiLSTM-based anomaly detection in multivariate time series with attention mechanism and dual analysis. In 2024 IEEE 7th International Conference on Information Systems and Computer Aided Education (ICISCAE) (pp. 379–384). IEEE. https://doi.org/10.1109/ICISCAE62304.2024.10761506
Article Sidebar
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Copyright Telecommunications Association Inc.
How to Cite
