Research On Equalization Technology Of Large-Capacity Optical Transmission System

In recent years,with the popularity of high-definition video,4K TV,mobile Internet and other services,the flow of information and communication networks has grown rapidly.However,the traditional backbone network cannot meet the increasing demands of users in the future,and it is urgent to expand and upgrade the communication system.In order to continuously increase the capacity of the optical transmission system,on the one hand,researchers use multi-dimensional parameters of light wave such as frequency,amplitude,phase,time slot and mode to carry information,and realize T-bit or even higher transmission rate.On the other hand,researchers continuously break through the rate limit of optical signal caused by crosstalk and nonlinear damage of fiber channel,and continuously increase the transmission capacity and transmission distance of the system under the existing signal dimension.However,both methods have higher requirements for digital signal processing algorithms,which not only ensure effective equalization of channel impairments,but also comprehensively consider the problem of algorithm complexity.Therefore,in high-speed and large-capacity optical communication system,it is of great significance to use relevant digital signal processing algorithm to equalize and compensate for channel impairments.On the basis of studying the theory of optical fiber communication system,this dissertation focuses on the nonlinear compensation algorithm based on feedback convergence suitable for high-speed and large-capacity optical fiber communication system and nonlinear compensation algorithm based on logistic regression,aiming at the problems of high complexity and low compensation accuracy of nonlinear compensation algorithm in longdistance coherent optical communication system.XT-MIMO-Volterra algorithm for multi-core fiber communication system is studied to solve the problem of signal transmission performance degradation caused by crosstalk and nonlinear device damage.In order to solve the problem of signal transmission performance deterioration caused by cross talk and nonlinear damage of components in low-mode fiber communication system,the joint training sequence algorithm of Jones matrix and Volterrabased nonlinear equalizer and compensation algorithm based on training sequence-assisted CMA are studied.The performance of the algorithm is verified on the thousand-kilometer coherent optical communication experimental platform,seven-core experimental platform and six-mode experimental platform respectively.The main work and innovations of this dissertation are as follows:1.Research on Nonlinear Compensation Algorithm in Long Distance Coherent Optical Communication System(1)To solve the problem of high complexity and low accuracy of nonlinear compensation algorithm in long-distance coherent optical communication system,a DBP compensation algorithm based on feedback convergence is proposed.The first-order perturbation term of the nonlinear Schrodinger equation is used to estimate the nonlinear noise of the current signal,and the step size update of the symmetric split-step fast Fourier transformed(SSFFT)and the fast convergence of the nonlinear noise tolerance are realized by threshold decision.The algorithm is experimentally verified on an 80Gbit/s polarization multiplexing 16QAM 1000km coherent optical transmission system.The experimental results show that compared with the traditional algorithm,the nonlinear compensation algorithm based on feedback convergence can obtain about 1.1 dB OSNR gain at the forward error correction coding threshold(3.8×10-3).(2)A DBP compensation algorithm based on logistic regression is proposed.The algorithm first estimates the time error,and the static hidden layer parameters such as polarization rotation and phase rotation on the training set,so that the logistic regression algorithm can be trained in the symbol layer.Then,the Mean Square Error(MSE)of the signal is used to measure the nonlinear damage in the transmission process.According to the error damage function of logistic regression,the nonlinear operator in DBP algorithm is modified to achieve nonlinear compensation.The algorithm is experimentally verified on an 80Gbit/s polarization multiplexing 16QAM 1000km coherent optical transmission system.The experimental results show that the nonlinear compensation algorithm based on logistic regression can obtain an OSNR gain of 0.6 dB,and the complexity is reduced by 19.2 percent and CPU running time is reduced by 19.6 percent on average compared with the traditional algorithm.2.Research on XT-MIMO-Volterra Algorithm in Seven-core Optical Fiber Transmission SystemIn order to solve the problem of the degradation of signal transmission performance caused by nonlinear damage of devices and inter-core crosstalk in the seven-core space division multiplexing optical communication system,the transmission characteristics of the seven-core optical fiber transmission system are studied and the XT-MIMO-Volterra algorithm is proposed in this dissertation.The algorithm extracts the time cumulation of IC-XT in the multi-core fiber and improves the butterfly structure in the MIMO-Volterra algorithm to compensate the damage of the multi-core fiber.The algorithm is experimentally verified on a 280Gbit/s KK16QAM 2.5km seven-core optical communication system.The experimental results show that compared with the SISO-Volterra algorithm,the proposed XT-MIMO-Volterra algorithm can achieve a CSPR improvement of 1.3 dB,and compared with the MIMO-Volterra algorithm,the proposed XT-MIMO-Volterra algorithm can achieve 0.6dB CSPR boost.3.Research on Compensation Algorithm for Few-mode Optical Communication System(1)In order to solve the problem of signal transmission performance deterioration caused by crosstalk and nonlinear device damage in fewmode fiber communication system,a joint training sequence algorithm Jones matrix and Volterra-based nonlinear equalizer is proposed.The algorithm obtains the Volterra polynomial coefficients by decomposing the nonlinear Schrodinger equation,and uses the training sequence to compensate for the nonlinear damage caused by the device.The algorithm is experimentally verified on a 20GBaud PAM4/PAM8 10km few-mode optical fiber communication system.Experimental results show that the proposed algorithm can effectively alleviate the nonlinear and dynamic crosstalk in FMF system.The Jones-Volterra training sequence can achieve a 0.5 dB gain in received power at the Forward Error Correction coding threshold(3.8×10-3).(2)A compensation algorithm based on training sequence-assisted CMA is proposed,and the performance of uniform and probabilistic shaping 8level pulse amplitude modulated signals on a 10km six-mode graded index fiber is studied.The algorithm improves the convergence algorithm in the CMA algorithm through the training sequence,overcomes the dynamic characteristics and modal coupling of the few-mode link,and improves the accuracy of error convergence.The experimental results show that TACMA can obtain about 1dB receiver sensitivity gain.