Research On Passive Optical Network Signal Processing Method

To cope with the growing demand for data services,passive optical network(PON)technology adopts a single-point-to-multipoint topology to realize many subscriber accesses at a low cost.However,with the rapid development of 5G/B5G mobile Internet technology and various new types of network services and application scenarios with diversified demands,the PON technology at the access side needs to cope with the enormous bandwidth and signal transmission quality challenges.Coherent PON technology using new multicarrier systems and signal processing methods has more potential for carrying high-speed services and is a significant development direction for future access networks.Among them,multicarrier generation technology and minimal coherent reception technology can significantly reduce the system cost of high-speed transmission;the use of advanced signal equalization,carrier rate matching methods,and other technologies can effectively improve the performance of coherent PON systems.In this thesis,we study the digital signal processing method in coherent PON and propose a frequency division multiplexing system scheme based on minimal coherent receiving,a frequency division multiplexing nonlinear equalization scheme based on the Gaussian mixing model,and a multi-signal frequency response adaptation convergence scheme based on recursive least squares multiple-input-multiple-output,and simulate and analyze the proposed scheme.BER of the PON system and increase the capacity of the access system,which enhances the reliability and transmission capacity of the optical communication system.The main research work of the thesis is as follows:(1)The thesis studied the orthogonal frequency division multiplexing(OFDM)technology and filter bank multicarrier technology(FBMC)after probability shaping(PS),as well as two frequency division multiplexing systems based on minimalist coherent receiving.Firstly,the PS-OFDM and PS-FBMC simulation systems are built,and the changes in BER,power,and spectral density before and after PS between OFDM and FBMC systems are compared.Verified that after adding PS,the bit error rate performance of the two systems showed a significant decrease,and the performance was effectively improved.Due to the more obvious spectrum leakage in the OFDM system,the system’s spectrum leakage has been reduced to a certain extent after PS.Then,a frequency division multiplexing scheme based on minimalist coherent reception is proposed,and the implementation process of PS-OFDM and PS-FBMC based on minimalist coherent reception is elaborated.A simulation system is built and verified.The results show that in the minimalist coherent reception system,the bit error rate still significantly decreases after adding probability shaping to the two frequency division multiplexing systems,but there is no significant difference compared to traditional coherent reception systems.The minimalist coherent reception technology does not significantly impact system performance while simplifying the complexity of four system components:OFDM,PS-OFDM,FBMC,and PS-FBMC.(2)A nonlinear equalization scheme for frequency division multiplexing based on the Gaussian mixture model(GMM)is proposed,effectively improving the performance of OFDM systems.Firstly,the simulation verified the optical frequency comb generation system based on a dual drive Mach Zender(MZM)modulator,cascaded with MZM and phase modulator(PM)and cascaded with electric absorption modulator(EAM)and frequency modulator(FM).An OFDM system based on an optical frequency comb was designed on this theoretical basis,and its performance was verified.A clustering algorithm based on GMM was studied to reduce nonlinear effects in OFDM systems.The GMM model can fit complex models with several Gaussian models and approximate all models.Subsequently,a simulation was conducted and compared with existing kmeans algorithms.Simulation analysis showed that using the GMM model can effectively improve performance.(3)A multi-signal frequency response adaptive convergence scheme based on Recursive Least Squares Multiple Input Multiple Output(RLSMIMO)was proposed and validated,which can flexibly allocate signals with different modulation formats according to the communication quality of different frequency bands,achieving optimization of overall communication performance.Upconversion aggregates signals transmitted by multiple users with different modulation formats and bandwidths.Allocate high-order modulation signals in the baseband or low-frequency frequency band.For signals using low-order modulation formats,allocate them in the high-frequency band to balance the transmission effects of different signals.Simulation shows that this scheme can simplify the aggregation layer structure and have better transmission performance.Compared with the random allocation scheme,the frequency response adaptation convergence scheme can increase the transmission distance of the system from 3km to 10km.Since the multi-input multi-output channel is also an essential factor in generating inter-symbol crosstalk,to reduce the decrease in communication quality caused by inter-symbol crosstalk,the RLS-MIMO algorithm is used to improve system performance.Comparing the bit error rates of PAM4,PAM8,and PAM16 signals before and after RLS-MIMO,it can be seen that RLS-MIMO has a significant impact on the bit error rate of PAM4 signals,and the bit error rate decreases significantly before and after RLS-MIMO.