A method and related equipment for eavesdropping classification monitoring based on recurrent neural network

Abstract

One or more embodiments of this specification provide a method for monitoring eavesdropping classification based on a recurrent neural network, including: collecting a transmission signal transmitted on a fiber channel, and determining parameters of the transmission signal; the parameters include an optical signal-to-noise ratio, an optical signal-to-noise ratio, a Dispersion, polarization mode dispersion, wavelength and bit error rate; do preprocessing to the parameters to obtain input values; use a pre-trained recurrent neural network to analyze and calculate the input values ​​to obtain analysis results; according to the analysis results, Interfering with the eavesdropping behavior; obtaining the real-time stress of the optical fiber, and determining the eavesdropping classification result according to the analysis result and the corresponding real-time stress of the optical fiber. The solution of the present disclosure is beneficial to eliminate errors caused by system changes and reduce the occurrence of false detections, and formulate a specific anti-eavesdropping method for a specific eavesdropping method, thereby improving the security of the optical communication system without affecting the normal transmission of signals. .

Description

technical field [0001] One or more embodiments of this specification relate to the technical field of wiretapping monitoring, and in particular, to a method and related equipment for wiretapping classification monitoring based on a recurrent neural network. Background technique [0002] Optical fiber communication is widely used in many real-life scenarios due to its large-capacity and high-speed transmission characteristics. Therefore, many criminals try to use various means to obtain information transmitted in optical fibers for profit. The eavesdropping methods can be divided into two categories: light leakage eavesdropping and coupling eavesdropping. Light leakage eavesdropping is to use the leaked optical signal to restore the original signal by destroying the optical fiber, including the optical fiber bending method and the V-groove incision method. Coupling eavesdropping is to couple the transmission fiber and the eavesdropping fiber by physical or chemical means, an...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the purpose of one or more embodiments of this specification is to propose a wiretapping classification monitoring method based on a recurrent neural network to solve the problem that the wiretapping means cannot be identified in the prior art, so that a specific wiretapping method cannot be formulated for a specific wiretapping method. The problem of anti-eavesdropping methods

Images