Article · Advances in Artificial Intelligence and Machine Learning · January 1, 2025
AI-Based OTDR Event Detection, Classification and Localization in Optical Communication Networks
Islambek Saymanov, Rikhsi Isaev, Saparniyaz Tursimuratov, Avazbek Jurabekov, Давлетова Холисахон Рахимджановна
Abstract
The article describes the research on the assessment of the operational state of fiber-optic communication lines in operation in the sharply continental climate of the Republic of Karakalpakstan.The study analyzed the operational state of the fiber-optic communication line by processing the OTDR trace curves obtained from the OTDR (Optical Time Domain Reflectometer) device.The MATLAB program was used as an environment for processing the OTDR trace curves obtained from the OTDR device.In this case, the OTDR trace curve was converted from the SOR format to the CSV format and processed.In order to improve the quality of the initial return signal to the OTDR trace curves, filtering and noise were reduced, and events on the line were identified.At the next stage, artificial intelligence tools were used.The k-means clustering algorithm was used for analysis.In fiber-optic communication lines, line segments are divided into groups as normal, faulty or anomalous depending on the attenuation and loss of the optical signal power.Also, signal drops above the average level were detected using the threshold method.This increased the efficiency of early fault detection.The research results showed that the proposed methods are effective in processing large volumes of OTDR data.They also increase the accuracy of fault detection and localization.This research method is convenient and easy to use in conditions where initial data for analysis and research are scarce, that is, difficult to control.In addition, this 4631
More by Islambek Saymanov
- Compressed-Domain Temporal Cues and Degradation-Aware Distillation for Lightweight Dynamic Facial Expression Recognition on Edge DevicesTechnologies · 2026
- Action-Unit-Constrained Graph Learning with Counterfactual Recourse for Auditable Facial Expression RecognitionComputers · 2026
- Three-Class Correspondence Reliability for Deformation-Preserving ICP: Controlled Synthetic Evaluation and a Cross-Domain Stress TestJournal of Imaging · 2026
- An EMC-Aware Multi-Objective Framework for Joint Radio-Resource and Transmit-Power Coordination in Ultra-Dense 5G/Beyond-5G NetworksJournal of Sensor and Actuator Networks · 2026
