Study On Key Technologies Of Optical Performance Monitoring For Elastic Optical Networks

With the rapid development of data services such as cloud computing,internet of things,artificial intelligence,5G mobile communications,etc.,the global internet traffic demand shows an exponential growth trend,which brings huge challenges to the optical fiber communication network as the cornerstone of information transmission,forcing it to towards the direction of high speed,long distance,low cost and reconfiguration.However,in the face of complex network structure and scarce spectrum resources in the future,the traditional solution of dealing with link damage by adding redundancy or manual compensation is no longer applicable.Therefore,in order to reduce network operating costs,maximize resource utilization and ensure optimal operation of high-speed optical transmission networks,continuous monitoring of various optical performance parameters is essential,while baud rate(BR),modulation format(MF),optical signal to noise ratio(OSNR),residual chromatic dispersion(R-CD),and residual differential group delay(R-DGD)are used for optical performance monitoring.As the key parameters in optical performance monitoring,they are the focus of research.This paper focuses on the key technologies of optical performance monitoring in elastic optical networks.The main research contents can be summarized as the two following aspects:1.We propose a scheme to simultaneously implement baud rate identification(BRI),modulation format identification(MFI)and multiparameter optical performance monitoring(OPM)for elastic optical network(EON).First,the scheme is based on the enhanced picture of Radon transform(EPRT)which contains more signal phase,amplitude and density information,and the EPRT obtained from the original signal constellation of elastic optical network is used as the key feature of signal parameter identification and monitoring after gray processing,edge detection and radon transformation.Subsequently,the EPRT is fed into a multi-task learning(MTL)neural network to simultaneously perform BRI,MFI,chromatic dispersion identification(CDI),differential group delay(DGD)monitoring,and OSNR estimation.The results show that the accuracy of BRI,MFI and CDI can reach almost 100% after 105 epochs,114 epochs and 174 epochs for 14/28 GBaud polarization division multiplexing(PDM)-EON transmission system,respectively.For the OSNR estimation task,its mean absolute error(MAE)converges to 0.468 d B after 187 epochs,and the MAE of DGD monitoring stabilizes at 0.021 symbol period after 170 epochs.Subsequently,we further verified the effectiveness of this scheme through a PDM-EON experimental system at 14/28 GBaud.The results show that the accuracies of MFI and BRI achieve almost 100% after 118 and 175 epochs,respectively.For the OSNR estimation task,its MAE converges to 0.542 d B after 200 epochs.2.A multi-parameter optical performance joint monitoring scheme of coherent optical transmission system based on probability shaping(PS)transmitter is proposed,including baud rate identification,modulation format identification,OSNR monitoring and probability shaping rate(PSR)estimation.The advantage of this solution is that asynchronous delay-tap picture ADTP is selected as the key feature of signal parameter identification and monitoring.Only 1-2 low-speed analog-to-digital converters and electrical delays can be used to monitor multiple parameters,and the hardware implementation cost is relatively low.The feasibility of the scheme is verified by PS-EON simulation transmission system.The simulation results show that BRI and MFI reach 94.6% and 99.8% identification accuracy after 179 and 170 epochs,respectively.For the OSNR estimation task,MAE converges to 0.21 d B after 187 epochs and the MAE for PSR monitoring stabilised at 0.21 bits/symbol after 170 epochs.