Research On Theoretical Models And Compensation Algorithms Of Fiber Nonlinearity In Coherent Optical Communication Systems

The fiber nonlinearity is a critical issue that limits the capacity of optical fiber communication systems.With the massive growth of the global communication services,the capacity of the deployed optical fiber communication system has approached the nonlinear Shannon limit,and it is important to study how to overcome the influence of the fiber nonlinearity and break through the nonlinear Shannon limit.The studies of the theoretical models and compensation or mitigation methods of the fiber nonlinearity attract much attention in recent years.The existing nonlinearity theoretical models are not comprehensive enough,and the existing fiber nonlinearity compensation methods need to further improve the performance and reduce the complexity.Based on the above background,this thesis studies the theoretical models and compensation algorithms of the fiber nonlinearity in bidirectional Raman amplified unrepeatered systems and wave division multiplexing(WDM)systems.Two fiber nonlinearity theoretical models are proposed,and based on the proposed theoretical models three fiber nonlinearity compensation algorithms are proposed.The major research achievements and contributions are summarized as follows:(1)A theoretical model of the bidirectional Raman amplified unrepeatered system is proposed.The proposed theoretical model includes not only the fiber nonlinearity caused by the signal itself but also the nonlinear signal-noise interaction(NSNI).The proposed model can accurately and quickly calculate the power of the nonlinear interference in the bidirectional Raman amplified unrepeatered system.To a certain extent,the proposed theoretical model makes up the defect that the existing model is not suitable for the bidirectional Raman amplified system.(2)Based on the above theoretical model,a nonlinearity compensation algorithm for bidirectional Raman amplified unrepeatered system is proposed,which combines the nonlinearity pre-and post-compensation.The proposed method can suppress the NSNI in the system and improve the performance of nonlinearity compensation.The simulation and experimental results show that the proposed method can effectively suppress the NSNI in the bidirectional Raman amplified system.Compared with the traditional digital back-propagation(DBP)algorithm,without increasing the complexity the proposed method makes more than 1 dB improvement over the DBP.(3)The influence of the fiber nonlinearity on the signal in WDM systems is studied,and a fiber nonlinearity analytical model is proposed.The proposed analytical model can analyze the influence forms of the fiber nonlinearity on the signal in different conditions and guide the design of fiber nonlinearity compensation methods,which can improve the performance of nonlinearity compensation.(4)Based on the proposed nonlinear analysis model,a joint compensation algorithm is proposed to mitigate the cross-phase modulation(XPM)in WDM systems.The proposed method first controls the interaction between the fiber nonlinearity and the dispersion by the nonlinearity pre-compensation,which makes the influence of the fiber nonlinearity on the signal more trend to the nonlinear phase noise.Then the phase recovery algorithm is used to compensate for the nonlinear phase noise at the receiver,and finally,the XPM effect is mitigated.The simulation results show that the proposed method can improve the signal-to-noise ratio(SNR)of the signal by about 0.4 dB in an 11-channel 1000-km transmission.Compared with other methods,the proposed method does not need the information of other channels and the algorithm is simpler and easier to implement.(5)A more efficient nonlinear phase noise tracking method based on the modified decision-directed recursive least squares(DD-RLS)algorithm is proposed.The modified DD-RLS algorithm has faster convergence speed and better noise tolerance,which can better track the nonlinear phase noise than the traditional DD-RLS algorithm and other phase recovery algorithms.Combining with the nonlinear pre-compensation the proposed modified DD-RLS algorithm can effectively mitigate the XPM in WDM systems.In the simulation,the proposed method can improve the Q~2-factor of the signal by 0.8 dB in an11-channel 1000-km transmission.The performance of this algorithm is better and the computational complexity is less than the existing algorithms.