Welcome to Smart Optical Fiber System Lab

News Center

Location:Home > News Center >Channel modelin...

Channel modeling of P2MP coherent PON system based on diffusion model

04
09
2025

Channel modeling of P2MP coherent PON system based on diffusion model

    Recently, Professor Yi Lilin's research group from the Department of Electronic Engineering at Shanghai Jiao Tong University (LIFE, Laboratory of Intelligent Fiber Ecosystem) studied point to multipoint (P2MP) coherent passive optical network (PON) systems and successfully applied the diffusion model to optical communication channel modeling for the first time. The effectiveness and practicality of the model were verified, and the accuracy and generalization ability of the model in P2MP coherent PON system channel modeling were deeply explored. The signal-to-noise ratio modeling error was less than 0.1dB in the range of 8dBm to 14dBm transmission power. The related results were published in the international optical journal Optics Letters in May 2025 under the title "P2MP coherent PON system modeling based on diffusion model". PhD students Guan Xiaokai and Xu Yongxin are co first authors, and Professor Yi Lilin is the corresponding author.

图1. 基于DSCM的P2MP下游架构示意图.png图2. 使用扩散模型进行通道建模的模型结构(a)训练过程.png图2. 使用扩散模型进行通道建模的模型结构(b)取样过程.png图 3 . (a)16QAM 信号的扩散过程(b)QPSK 、16QAM 和64QAM 信号的去噪处理.png图4. 扩散模型的通信建模结果(a)无非线性(b)有非线性(c)SNR曲线和符号速率.png图5. 通道建模结果(a)相同配置的4个ONU (b)不同发射功率下的ONU-1.png图6. 通道建模结果 (a)不同发射功率下的四个不同ONU (b)不同发射功率下不同光纤长度的ONU-1 (c)不同链路长度下不同 ONU数量的ONU-1.png