Research On Modulation Format Identification Technology For Probabilistically Shaped Signals

Probabilistic shaping combined with coherent detection technology can adaptively adjust the data rate,bandwidth,wavelength and other system resources according to the communication link conditions,system resources,etc.to maximize the utilization of system resources and ensure the quality of service,which has been emerging as one of the key techniques for next intelligent optical communication networks.The identification techniques for modulation format and shaping distribution information can provide the basis informanction to the modulation format dependent algorithms at the receiver,so as to reduce the complexity of the receiver and acheive the optimization performance.Based on this,this thesis studies the modulation format and shaping distribution information identification technology for probabilistic shaping systems.The main acheivement are listed as following:1.A signal distribution based modulation format identification shceme for probabilistically shaped M quadrature amplitude modulation(PS-MQAM)is studied.Firstly,the cumulative intensity of the probability shaping signal in two specific regions is calculated,and then the cumulative intensity in the two regions is mapped to a two-dimensional plane.Finally,PS-MQAM shaped modulation format indentification is obtained by utilizing the the support vector machine(SVM)algorithm.A 28 GBaud PDM-PS-16QAM/-32QAM/-64QAM/-128 QMA probability shaping coherent optical fiber communication simulation system is built up to verified the feasibility of the scheme.The results show that the four PSMQAM modulation formats can achieve 100% recognition accuracy when optical signal-tonoise ratio(OSNR)is less than normalized generalized mutual information(NGMI)threshold of 0.935.2.According to the nonlinear power variation characteristics of the PS-MQAM probability shaping signal,a modulation format identification scheme based on the nonlinear power variation of the signal is proposed.Firstly,the amplitude of the probability shaping signal and the characteristic information related to the modulation format in the signal power spectrum after the signal has been passed to the M power(4,8,16)are calculated,and then the DNN is used to extract the relevant characteristic information to complete the identification of the modulation format information.The simulation results show that in the 28 Gbaud PDMPS-16QAM/-32QAM/-64QAM/-128 QMA signal mixed with various shaping parameters,the method can still achieve 100% modulation when the OSNR value is lower than the NGMI threshold requirement of 0.935 format recognition accuracy.3.According to the amplitude distribution characteristics of the PS-MQAM probability shaped signal,a shaping distribution identification scheme based on the signal amplitude characteristics is proposed.First,by calculating the amplitude and variance of the normalized signal and the number of symbols in each amplitude region,the characteristic parameters related to the shaping distribution are obtained.Then,DNN is used to extract all the characteristic parameter information related to the shaping distribution,so as to realize the shaping distribution identification of the specific modulation format.The simulation results show that in the 28 GBaud PDM-PS-16QAM/-32QAM/-64QAM/-128 QMA back-to-back coherent optical communication system,the OSNR can achieve a constellation probability distribution entropy interval of 0.2bit/ symbol 100% recognition.PDM-PS-64 QAM signal can realize 100% identification with constellation probability distribution entropy interval of0.1bit/symbol when the fiber input power is 0d Bm in 800 km SSMF transmission.