Neural Network System for Monitoring State of a High-Speed Fiber-Optical Linear Path

摘要

The paper presents a methodology for the synthesis of systems for monitoring the state of a high-speed fiber-optical linear path, based on the information and measuring control system that implements neural network recognition algorithms with synthesis by dominance. The proposed information-measuring system processes levels of the average intensity of the optical signal received at various carriers in a certain retrospective over a defined period of time. Neural network algorithms are synthesized by optimizing the parameters of a neural network that allow to make a conclusion about the state of a high-speed fiber-optical linear path.