新型互联网业务的出现促使光通信系统朝着高速、大容量、长距离、灵活的方向发展，为了保证光通信系统的良好运行，作为关键技术的光性能监测则变得更为重要。传统的监测技术仅仅依靠OSNR和色散监测即可保证系统良好运行，然而，随着光通信的发展，非线性效应已经成为限制光通信发展的主要因素，非线性效应影响下的OSNR和色散的监测性能也难以满足未来光网络传输的要求；光网络的灵活性趋势对光性能监测提出了更高的要求。本项目拟通过研究光通信网络的非线性效应、OSNR和色散的作用机制，建立高精度的多因素损伤模型；在此基础上，设计特征统计数据和深度学习网络架构，不仅可以实现灵活光网络调制格式和传输速率的透明与识别，还可以实现非线性损伤、OSNR和色散的高精度多任务监测；提出多输入-多输出的深度学习网络和迁移学习方案，实现大容量光通信系统的多信道并行监测。本项目的研究为未来光通信系统的智能监测和应用奠定了坚实的基础。

The development of new Internet services promotes the optical network system towards the direction of high speed, large capacity, flexibility and long distance. In order to ensure the transmission performance of optical communication system, optical performance monitoring is one of the essential key technologies. The traditional monitoring technology only needs and realizes OSNR and dispersion monitoring to ensure the operation of the system. However, with the development of optical communication, the nonlinear effect has become the main factor limiting capacity of optical communication. The OSNR and dispersion monitoring technology under the influence of the nonlinear effect are also difficult to meet the requirements of future optical network transmission. In addition, the characteristics of flexible optical networks require the monitoring technology to be transparent of the modulation format and rate information, which also improves the requirements for optical performance monitoring. This project intends to establish a high-precision multi-factors damage model by studying the interaction mechanism of nonlinear effect, OSNR and dispersion for this system. Based on this model, the design of feature statistics and deep learning network architecture not only realizes identifications of modulation format and transmission rate, but also realizes the high-precision multi-task monitoring with nonlinear damage, OSNR and dispersion. A multi-input-multi-output deep learning network and transfer learning technology are established to realize multi-channel parallel monitoring of high-capacity optical communication system. The research of this project establishes a solid foundation for the intelligent monitoring and application of future ultra-high capacity optical communication system.