Research On Key Parameters Estimation For Optical Fiber Signal Monitoring

With the increasing demand for various high-bandwidth services,the transmission rate of optical fiber signals is doubling every year,a variety of related technologies are also constantly breakthrough innovation,such as single-mode fiber,single-frequency laser and multiples of advanced modulation formats.On the other hand,in order to meet the demand of cost saving and network flexibility,the data path and rate of transmission in a fiber-optic network are not fixed.And the fiber signal channel will decay and ultimately key parameters change due to temperature,device parameter drift,aging and other reasons.Typical optical fiber signal key parameters include: symbol rate,modulation format,transmission signal quality evaluation parameter-optical signal to noise ratio,chromatic dispersion and polarization mode dispersion.The accurate estimation of the key parameters of optical fiber signal can enable the node in the complex fiber network to distribute the bandwidth resource and diagnose the fault error effectively,and finally realize the intelligent management of the optical fiber network.From the perspective of national security,fiber-optic network is an important part of cyberspace,which is the another battlefield following the sea,land,air and universe.Optical fiber signal monitoring is an important method to realize fiber network situation perception,fiber signal key parameter estimation and modulation format identification are the main content of fiber signal monitoring.Especially in non-collaborative conditions,they are also complementary and mutually influential.In this paper,we study the key parameter estimation and modulation format identification of optical fiber signal,the main contribution and novelty are listed as follow:(1)A symbol rate estimation method of optical fiber signal based on asynchronous delay tap sampling and average amplitude difference function is proposed.Existing fiber signal symbol rate estimation techniques are mostly aimed at known signal prior information,the sampling rate requirements are relatively high,thus it cannot satisfy the demand of symbol rate accurate estimation.Especially the additional amplitude modulation optical fiber signals,which need to consume more physical logic resources during the symbol rate estimation prcocess.To realize the efficient estimation of the symbol rate of optical fiber signal,the use of asynchronous delay tap sampling rate is much lower than the symbol rate,By adjusting the delay time of asynchronous delay sampling,the average amplitude difference function is obtained for two asynchronous sampling data.By observing the periodic characteristic value of the average amplitude difference function,the reciprocal of the value is the estimation symbol rate of the optical fiber signal.(2)A modulation format identification method of optical fiber signal based on asynchronous delay tap plot entropy and support vector machine is proposed.In order to save resources and improve the efficiency use of bandwidth in the optical fiber network,multiple of modulation formats and rates are simultaneously utilized and then the complexity of fiber signal parameters are increased,the identification of the modulation format for useful information in a particular wavelength is a critical step in fiber signal acquisition and monitoring.Based on the previous proposed asynchronous delay tap two-dimensional plot as the modulation format identification feature,the depth information extraction is carried out.A new four-dimensional eigenvector is proposed,which is consisted of shannon entropy,exponential entropy,singular shannon entropy and singular exponent entropy based on the entropy characteristic.The increase of the feature vector dimension can improve the identification accuracy.At the same time,the support vector machine is applied with the advantage of the small sample learning requirements for data classification and no local optimal solution among global solution.Real-time modulation format identification is realized for high-speed optical fiber signal under the condition of limited hardware resources,and the method is proved to be compatible with the new advanced modulation format.(3)The optimal OSNR estimation method is proposed for multiple modulation formats.As an important fiber signal parameters,optical signal-to-noise ratio(OSNR)is frequently used for fiber-optic connection fault diagnosis,optical fiber signal quality assessment and reconfigurable fiber-optic network monitoring.According to the characteristics corresponding to respective signal modulation formats,the optimal statistical region is studied by using the asynchronous delay low-speed sampling technique,and the statistical distribution characteristic optimization of the sampling data is achieved through the two-dimensional distribution plot of the sampling data,so as to obtain the optimal OSNR.Especially for the additional amplitude modulation optical fiber signal,an OSNR estimation method based on peak-to-valley ratio and high-order polynomial fitting is proposed.The maximum error is only 0.8dB.In addition,the method can be used with the tunable dispersion module for blind chromatic dispersion compensation in optical domain.(4)For the different optical signal modulation formats,a blind chromatic dispersion estimation method using the maxing peak-to-valley ratio of average amplitude difference function is proposed in the optical domain.Additionally,an improved chromatic dispersion compensation method based on coherent detection is proposed in the electronic domain.Chromatic dispersion is the main cause of inter-symbol interference(ISI)for high-speed optical fiber signal.Traditional CD estimation method is based on pre-determination of fiber signal transmission distance in cooperative condition.In the flexible reconfigurable fiber network and non-cooperative communication conditions,the transmission distance may be continuously changing and unknown.In order to solve the problem of blind CD estimation,the peak-to-valley ratio of the amplitude difference function is obtained by using the asynchronous delay tap sampling for different modulation formats.The maximum peak-to-valley ratio is obtained through the dispersion tunable module scanning,and the final adjustment value is the estimated CD value.For the advanced modulated optical fiber signals,such as PM-DQPSK,we can use autocorrelation of signal power waveform(ACSPW)to roughly estimate chromatic dispersion and obtain a small scanning range.Then singular value of signal autocorrelation matrix(SVSACM)is utilized to scan in the estimated range for accurate CD estimation and the rate extension?This method can effectively extend the ACSPW symbol rate limitation,while it can also enhance the SVSACM efficiency by narrowing the CD searching range.