Research On The Technology Of Modulation Format Identification Based On Clustering Analysis Method In High-speed Optical Fiber Communication System System

Along with the high-speed development of machine learning technologies,and driven by emerging applications such as cloud-based flexible broadband services,IP video and wireless backhaul,optical networks are undergoing tremendous changes.Based on the development needs of optical network networking,scheduling,operation,etc.,intelligence has become one of the key development directions for future optical networks,however,the achievement of future flexible and elastic optical network architecture requires extremely complex management.In order to implement flexible network management,the intelligent cognitive optical network(ICON)grew up,in which the rate-adaptive transceiver supports multiple modulation formats,has good scalability and collaboration,and can coordinate transmitter and receiver to achieve realtime dynamic adjustment of optical signal rate and modulation format,and realize link management functions such as intelligent channel management and bandwidth allocation of network nodes.At the same time,the signal quality in the optical communication network is monitored by optical performance monitoring(OPM)technology,which provides the basis for dynamic routing and network management.In most cases,the abovementioned functions of intelligent optical network management,OPM and optical receivers require the identification of information such as the symbol rate and modulation format of the optical signals first,therefore,modulation format identification(MFI)technology is important for the next-generation optical fiber networks,because it provides more autonomy and flexibility for the network.This paper focuses on MFI in ICONs.Taking the constellation diagram of the signals in Jones space as the feature,an MFI scheme based on non-iterative clustering method and Calinski-Harabaz(CH)index is proposed.At the same time,in terms of alleviating signal impairments,a pre-compensation scheme is proposed to deal with some common signal impairments.The following is the main work that has been done in this paper:1.Through investigating the compensation methods for common signal impairments,a pre-compensation scheme integrated by frequency offset estimation by fast fourier transform(FFT-FOE)and blind phase search(BPS)for 4QAM is proposed in this paper.2.An MFI scheme based on non-iterative clustering method and CH index is proposed.The scheme can respond to the real-time demands of MFI in the optical communication system,and has the merits that easy to achieve and low-cost.3.For the MFI scheme proposed in this paper,use OptiSystem and Python to build a 28 Gbaud polarization division multiplexing(PDM)optical fiber communication system simulation experiment platform to test the performance of the scheme.For PDM-QPSK,PDM-16QAM,PDM32QAM and PDM-64QAM,a total of 4 different modulation formats of the signals,the simulation results show that the scheme can achieve 100%identification accuracy even when the optical signal-to-noise ratio(OSNR)is lower than or equal to the OSNR corresponding to the 7%forward error correction(FEC)threshold of each modulation format.Also,the scheme can also perform well when residual dispersion exists in the simulation system,and the number of symbols required for the scheme to achieve 100%recognition accuracy is also low.At the same time,through the comparison with other scholars’ research,the high efficiency and stability of the scheme proposed in this paper is fully demonstrated.In addition,for the sake of evaluating the performance of the scheme further,a 20 Gbaud PDM longhaul optical fiber transmission experiment is built in this paper to test the identification accuracy of the scheme for PDM-QPSK,PDM-16QAM and PDM-32QAM at different transmission distances and different launch powers.from the results of the experiments,it is easy to find that the scheme has a high tolerance for nonlinear impairments.