本项目旨针对未来卫星激光通信网络的高速率、低时延、大容量、高可靠等要求，对高速的卫星激光通信系统的链路性能优化方法进行研究，结合机器学习理论，提出链路性能优化方案，为实现卫星激光链路的高速数据传输的工程应用提供理论技术基础。主要研究内容包括：1）建立多普勒频移评估模型，提出基于EM算法的多普勒频移评估补偿方案；2）基于改进型最小二乘法的带内串扰抑制方案研究；3）分析背景光噪声对卫星激光通信系统的影响，提出基于K-means算法的背景光噪声抑制方案；最后搭建高速基于波分复用的卫星激光通信系统仿真平台，验证链路性能优化方案的可行性。通过本项目的研究，期望在空间激光通信这一技术研究方向取得扎实的进展，为高速卫星激光通信的工程实现积累成果。

The purpose of this project is to research the link performance optimization method of high-speed satellite laser communication system, aiming at the requirements of high-speed, low-delay, high-capacity and high reliability of future satellite laser communication network. Combining machine learning theory, we propose the optimization scheme of link performance based on theoretical and technically in satellite communication. The main research contents are as follows: 1) Establishment of Doppler frequency shift evaluation and compensation model, proposed Doppler frequency compensation scheme based on EM algorithm; 2) We propose in-band crosstalk suppression scheme based on advanced least squares method; 3) The effect modeling of of background light noise on the satellite laser communication system is proposed, and the background light noise suppression scheme based on K-means algorithm is proposed. Finally, a high-speed satellite laser communication system platform based on wavelength division multiplexing is proposed to verify the feasibility of the link performance optimization scheme. Through the research of this project, it is expected to make solid progress in the field of laser communication technology, and to accumulate the research results for high-speed satellite laser communication engineering.