Machine Learning Based Denoising Anomaly Detection and Localisation Using BiGRU in Optical Fiber Monitoring

Efficient data transmission in optical fiber networks demands robust fault detection and classification. Leveraging Denoising Convolutional Autoencoder (DCAE) for fault detection and Bidirectional Gated Recurrent Unit (BiGRU) for classification, our method achieves a 96% efficiency rate. By accurately identifying anomalies like bad splices, dirty connectors, cuts, and eavesdropping, it ensures network integrity and confidentiality. This approach mitigates financial losses and potential disruptions caused by signal loss and interference. Overall, it provides a reliable solution for safe and steady communication in high- bandwidth optical networks.