基于光子神经网络的光通信网络窃听定位

Eavesdropping Location in Optical Communication Networks Using Optical Neural Networks

摘要:

光通信网络安全至关重要，其中窃听行为的存在更是威胁系统的信息传递。为了解决这个问题，提出了一种基于光子储备池计算(photonic reservoir computing, PRC)的简单高效窃听定位方法。通过仿真搭建了正交频分复用(orthogonal frequency division multiplexing, OFDM)光纤通信系统和基于光反馈半导体激光器的储备池结构。通过在光纤链路中引入光耦合器来模拟窃听行为，并在接收端采集相关性能监测数据。利用接收到的性能监测数据来训练光子储备池计算模型。在研究了一系列关键的半导体激光器参数后，仿真结果表明，该方法对窃听行为的位置识别达到了超过99%的识别准确率。该方法为使用简单的硬件实现光通信网络窃听定位提供了一条有前途的途径。

Abstract:

The security of optical communication networks is crucial, and the existence of eavesdropping poses a significant threat to the transmission of information within the system. To address this issue, a simple and efficient method for eavesdropping location based on photonic reservoir computing (PRC) is proposed. An optical fiber communication system based on orthogonal frequency division multiplexing (OFDM) and a PRC structure based on optical feedback semiconductor lasers are constructed through simulation. The eavesdropping behavior is simulated by introducing a coupler in the optical fiber link, and the relevant performance monitoring data are collected at the receiver. The received performance monitoring data are utilized to train the PRC model. After optimizing a series of key parameters of semiconductor lasers, simulation results show that our approach achieves more than 99% recognition accuracy for eavesdropping location. This method provides a promising way to implement eavesdropping location in optical communication networks using simple hardware.