Volterra Adaptive Filter for Real-Time OTDR Monitoring and Fault Localisation in Fibre Optic Networks

Abstract

Locating a fault in a fibre-optic cable is often comparable to searching for a needle in a haystack of noisy backscattered signals. To address this challenge, this paper presents a Volterra Adaptive Filter (VAF) framework for enhanced Optical Time Domain Reflectometry (OTDR)–based monitoring and fault detection in fibre-optic networks. The proposed approach is centred on a first-order Volterra adaptive filter with optimised memory length and momentum-stabilised learning, designed to improve prediction stability and tracking accuracy under noisy conditions. Comprehensive experimental validation was conducted using a 20 km operational fibre route in Lagos, Nigeria. The system achieved a prediction root-mean-square error of 0.759 dB, a mean absolute error of 0.064 dB, and a correlation coefficient of 0.9974, indicating excellent agreement with measured OTDR traces. An integrated multi-criteria fault detection scheme attained 80 % precision and 80 % recall, corresponding to an F1-score of 0.800, while processing 19 980 samples using only 20 adaptive parameters. The proposed system localises faults to specific, named network components and translates complex reflectometry signatures into actionable maintenance information for field personnel. Operating in real time at processing rates exceeding 1 500 samples per second, the framework demonstrates the robustness and computational efficiency required for deployment in live fibre-optic network operations.