On January 6, 2023, the optical communication simulation software jointly sponsored by the Optical Communication Committee of the Chinese Society of Communications, Pengcheng Laboratory, the State Key Laboratory of Regional Optical Fiber Communication Networks and New Optical Communication Systems, and the School of Electronic Information and Electrical Engineering of Shanghai Jiao Tong University The development seminar and the conference of the open source simulation platform for intelligent optical transmission were successfully held online and achieved complete success. A total of 328 members of the Optical Communication Committee of the China Society of Communications, some members of the State Key Laboratory of Regional Optical Fiber Communication Networks and New Optical Communication Systems, and experts in the optical communication industry participated in this conference online, and nearly 5,000 people were online in the Kouxiang Academic Live Room watch.

Intelligent fiber transmission simulation (IFTS), an open source simulation platform for intelligent optical transmission developed by the intelligent optical fiber communication system team of Professor Lilin Yi of the Department of Electronic Engineering of Shanghai Jiaotong University (LIFE, Laboratory of intelligent fiber ecosystem), was officially released at this conference. All codes are freely available to peers online.

Research Background

In the era of big data, the massive growth of data has brought opportunities and challenges to the development of optical communication. However, the research on optical communication relies on expensive experimental equipment and good experimental conditions. The threshold for experimental research is high, and only a few units can carry out it, which limits the development of the industry to a certain extent. In terms of optical communication system simulation, it mainly relies on foreign simulation tools, and there are risks of control. Even the simulation tools commonly used abroad have the problems of slow simulation speed and long time-consuming, which affect the research progress. At present, artificial intelligence and optical communication are deeply integrated. Intelligent analysis is an important trend in the evolution of optical communication simulation. Open source, intelligent, accurate, and fast optical communication simulation software is a common demand in the industry.

In this context, the intelligent optical fiber communication system team of Professor Lilin Yi's research group lasted three years, with more than a hundred references, and developed a Python-based open source simulation platform for intelligent optical transmission by integrating various algorithm libraries. Formal release at the development workshop. The intelligent optical fiber communication system team advocates openness and sharing, and is willing to work with peers to make domestic optical communication simulation tools bigger and stronger, promote the development of simulation software for China's optical communication industry, and contribute collective wisdom and strength to the development of China's optical communication industry.

Group photo of some participants

IFTS Features and Performance

IFTS is an open source simulation tool for intelligent optical communication based on Python, designed to meet the needs of scientists, optical communication engineers and researchers in related fields for optical communication simulation tools. The platform includes functions such as signal encoding and modulation at the sending end, fiber channel transmission, demodulation and decoding at the receiving end, and provides a Pytorch-based simulation version with classic and AI dual modes, which allows users to use GPU for simulation acceleration, and naturally supports optical Intelligent algorithm design for communication.

Figure 1 (a) VINN structure; (b) cosine annealing learning rate curve with hot restart; (c) loss curve

IFTS Key Features

Compared with the existing mainstream simulation software for optical communication, its main advantage is that both transceiver and transmission support and embed AI algorithms, which can realize fast and accurate simulation, and the underlying code is free and open source. The AI model in the simulation platform includes both a general AI algorithm library and an AI-based channel acceleration model. The AI algorithm library can support users to design their own AI algorithms and quickly verify them in the simulation system. Facing the pain point that the simulation speed of traditional modeling algorithms is too slow, the AI channel acceleration model of IFTS provides a solution for fast and accurate channel simulation. In multiple scenarios, the running time of the AI channel model is only 1% of the traditional modeling algorithm. Under the definition of different devices and different data lengths, the degree of time optimization of the AI model is different. In the best case, the AI channel model only takes one-thousandth of the time of the traditional modeling algorithm.

Figure 2 (a) Experimental setup of 100Gbps IMDD PON system; (b) Frequency response of back-to-back system; (c ) The eye diagram of the received signal before equalization; (d) The eye diagram of the received signal after equalization

AI algorithm module of IFTS

Figure 3  12dBm transmit power 100Gbps O-band PAM4 signal sensitivity curve

IFTS AI channel module

Currently, IFTS can be used by colleges and universities for teaching and scientific research, and can also be used by optical communication companies for system design and performance optimization. Since the pre-release in October 2022, more than ten units including China Telecom, Huawei, FiberHome, Pengcheng Laboratory, Beijing University of Posts and Telecommunications, and Huazhong University of Science and Technology have downloaded and used it. The team of Dr. Li Junjie, chief expert of China Telecom's optical transmission field tried the IFTS platform, and commented in his report that "IFTS is a milestone product for the localization of optical communication simulation technology." In the future, the IFTS plan will continue to be updated, complete the device effect simulation, increase the FEC codec module, establish a complete AI algorithm library, and realize the digital twin model of the experimental system and other phased goals. In addition, IFTS will further improve the online community and establish a platform for feedback and communication. In the end, it is planned to create a powerful domestic optical communication open source simulation platform, which can be used by the whole industry for free, and contribute to solving the "stuck neck" problem of optical communication simulation software.

LIFE optical transmission direction:
LIFE research group intelligent optical fiber communication system team conducts continuous and in-depth research on fast and accurate modeling of optical fiber channels and nonlinear compensation issues, and is committed to designing AI architecture and algorithms for optical fiber communication systems to achieve The goal of approaching the Shannon capacity limit of the channel. At present, the research group conducts in-depth research on intelligent modeling and simulation of optical fiber communication systems, fast and accurate modeling based on multi-channel high-speed long-distance optical fiber channels, end-to-end nonlinear compensation and global optimization of optical fiber communication systems, and development of intelligent fiber optic communication systems based on Python. The simulation platform contributes to the sustainable development of the field of intelligent optical communication. The work in this direction is supported by (National Natural Science Foundation of China Outstanding Youth Fund (62025503), Ministry of Science and Technology Key Research and Development Program (2019YFB1803803), and Shanghai Jiaotong University-Huawei Advanced Optical Technology Joint Laboratory).

To obtain IFTS open source code, please visit the official website
Official website address: ifibertrans.sjtu.edu.cn
Conference playback address: https://www.koushare.com/lives/room/837693