An Intelligent Alarm Method for Optical Fiber Network Based on Backtracking Single Alarm Information

作者：Haisu Zhang, Yibo Liu, Long Wang · 年份：2020 · DOI：10.1109/iwecai50956.2020.00018 · 被引用次数：1 · 研究领域：Advanced Fiber Optic Sensors、Advanced Photonic Communication Systems、Software System Performance and Reliability

Optical fiber transmission network has gradually become the mainstream communication transmission medium and the development of artificial intel-ligence provides new ideas for the prediction of fiber transmission network failures. Different from traditional method which using the alarm infor-mation only for troubleshooting after a fault has occurred, this paper applies the real-time and serialized characteristics of alarm information to predict fault, achieving that a single alarm message can be used to indicate the pres-ence or absence of a fault. The specific steps of this paper are as follows, first extracting the alarm feature by collecting the alarm information and malfunction information from optical fiber network, then building a smart alarm learner by using the machine learning algorithm, finally achieving the goal to predict the probability of a failure occurring in the future. This meth-od can be integrated into the optical fiber transmission network operation, and lay the foundation for further research on Intelligent alarm.