Physical layer secrecy method and system based on mutual information estimation neural network

Abstract

The invention provides a physical layer secrecy method and system based on a mutual information estimation neural network, and the method comprises the steps: S1, carrying out the training of an encoder and a decoder based on mutual information estimation and end-to-end deep learning, and obtaining a trained encoder and a trained decoder; and S2, physical layer confidential transmission is carried out by using the trained encoder and the trained decoder. According to the method, mutual information estimation and end-to-end deep learning methods are introduced, a confidentiality system for minimizing the mutual information of eavesdropper channels is realized through mutual information estimation, a reliable communication system for realizing the maximum mutual information of legal user channels is realized through end-to-end deep learning, the scheme gives consideration to confidentiality and reliability, and the method is suitable for popularization and application. The method is suitable for a long-distance high-speed optical communication transmission system.

Description

technical field [0001] The present invention relates to the technical field of optical fiber security communication system, in particular, to a physical layer security method and system based on mutual information estimation neural network, and more specifically, to a physical layer based on mutual information estimation and end-to-end deep learning Security method and system. Background technique [0002] In the field of physical layer security of optical fiber communication, traditional methods include XOR logic encryption, electrical logic encryption, and chaotic encryption. Synchronization, the added noise can be removed. However, these methods need to add complex optical devices to complete encryption and synchronization, and the encryption noise may be difficult to completely remove through channel damage, and synchronization errors may occur, resulting in an increase in the bit error rate at the receiving end; and these methods are not at the level of information the...

AI Technical Summary

Problems solved by technology

However, these methods need to add complex optical devices to complete encryption and synchronization, and the encrypted noise may be difficult to completely remove through channel damage, and synchronization errors may occur, resulting in an increase in the bit error rate at the receiving end; and these methods are not at the level of information theory Prove that the encryption method is the least informative against an eavesdropper and therefore at risk of being attacked

Images