Optical Fiber Signal Simulation Based On Machine Learning

In recent years,the application of machine learning technology and deep learning technology in life,learning and even in various fields has been deepened and improved.In the field of optical communication,its development is particularly rapid.In the research,artificial intelligence is of great significance to the optical communication.The combination of the two is not only limited to software,such as signal performance detection,signal processing,etc.,but also in hardware,as intelligent manufacturing of hardware,and intelligent monitoring and operation of the whole system will also be the focus of future development.In the aspect of the construction of optical fiber communication system,we always use the actual operation or professional software to simulate,which requires a large number of professional personnel to operate,and may cause some unnecessary material loss.In this paper,under the idea of saving resources and the cross combination of traditional industry and emerging intelligent fields,machine learning and deep learning are introduced into the simulation analysis of optical channel and the adaptive prediction of the signal at the receiving end of the optical communication system.The main work of this paper is as follows:First,the current research status and trend of simulation analysis of optical fiber channel and optical communication system are analyzed,and the importance and necessity of applying artificial intelligence algorithm to optical communication field are summarized.Secondly,combining with the typical cases of the application of artificial intelligence technology in optical communication,the advantages and disadvantages of traditional numerical method and artificial intelligence method in the simulation analysis of optical communication system are compared,and the applicable scenarios and application points of artificial intelligence algorithm in the research of optical communication technology are given.Thirdly,a fiber channel simulation scheme based on deep learning algorithms such as Decentralized News Network(DNN)and Long Short-Term Memory(LSTM)is designed and implemented,and the feasibility and accuracy of the scheme are analyzed.