基于峰均值功率比的色散估计算法研究

A Chromatic Dispersion Estimation Scheme based on the Peak to Average Power Ratio

针对下一代动态光网络,光纤长度可能根据链路需要随时变化,以及由于光纤老化造成色散参量变化等因素可能造成链路中的色散值在接收端很难提前预知,因此要求未来光网络节点中的接收端能够自适应地估计出色散值然后补偿信号。文章在光网络节点的数字相干接收机中提出了一种基于峰均值功率比的色散估计算法,并在224 Gbit/s偏振复用(PDM)-16正交振幅调制(QAM)和336 Gbit/s PDM-64QAM的多种长距离传输(400、800、1200、1600和2000 km标准单模光纤(SSMF))系统中验证了其可行性及有效性。仿真结果表明,文章所提算法具有很好的色散估计精度,色散估计标准差低于100 ps/nm。该算法对偏振模色散(PDM)和偏振相关损耗(PDL)具有较强的容忍性,例如群速度时延(DGD)低于15 ps及PDL在7 dB以内时,算法都可实现很好的估计精度。同时该算法对非线性效应也具有较强的容忍性。

For the next generation of dynamic optical networks,the fiber length of the transmission link may change dynamically according to the requirement of the optical networks and the chromatic dispersion coefficient may also change due to the aging of optical fiber,which result in the difficulty in the prediction of the cumulative chromatic dispersion value in the transmission link at the receiver side.Thus,it is required that the receiver in the future optical network node can adaptively estimate the cumulative chromatic dispersion for the future compensation process.In this paper,a chromatic dispersion estimation scheme based on the peak to average power ratio is proposed in the digital coherent receiver of optical network nodes.The proposed scheme has been demonstrated in a series of long-haul transmission links of 224 Gbit/s Polarization Division Multiplexing(PDM)-16 Quadrature Amplitude Modulation(QAM)and 336 Gbit/s PDM-64 QAM systems with 400、800、1200、1600 and 2000 km Standard Single Mode Fiber(SSMF).The results show that the proposed scheme can achieve good dispersion estimation accuracy(the standard deviation is less than 100 ps/nm).Meanwhile,the proposed scheme has good tolerance for Polarization Mode Dispersion(PMD)and Polarization Dependent Loss(PDL)(when the Differential Group Delay(DGD)is less than 15 ps and the PDL is less than 7 dB,the proposed scheme can achieve good estimation accuracy.).In addition,the proposed scheme can achieve chromatic dispersion estimation by using only one component of a polarization signal,which greatly saves the computational resources.In addition,the proposed scheme also has good tolerance for the fiber nonlinearity effect.