WO2020108632A1 - 色散估计方法、装置、接收机及存储介质 - Google Patents

色散估计方法、装置、接收机及存储介质 Download PDF

Info

Publication number
WO2020108632A1
WO2020108632A1 PCT/CN2019/122107 CN2019122107W WO2020108632A1 WO 2020108632 A1 WO2020108632 A1 WO 2020108632A1 CN 2019122107 W CN2019122107 W CN 2019122107W WO 2020108632 A1 WO2020108632 A1 WO 2020108632A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
phase angle
value
target
angle value
Prior art date
Application number
PCT/CN2019/122107
Other languages
English (en)
French (fr)
Inventor
李运鹏
Original Assignee
深圳市中兴微电子技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市中兴微电子技术有限公司 filed Critical 深圳市中兴微电子技术有限公司
Priority to JP2021531064A priority Critical patent/JP7482128B2/ja
Priority to EP19888798.6A priority patent/EP3890211A4/en
Priority to KR1020217019935A priority patent/KR20210096191A/ko
Publication of WO2020108632A1 publication Critical patent/WO2020108632A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/25—Arrangements specific to fibre transmission
    • H04B10/2507—Arrangements specific to fibre transmission for the reduction or elimination of distortion or dispersion
    • H04B10/2513—Arrangements specific to fibre transmission for the reduction or elimination of distortion or dispersion due to chromatic dispersion
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
    • H04B10/60—Receivers
    • H04B10/61—Coherent receivers
    • H04B10/616—Details of the electronic signal processing in coherent optical receivers
    • H04B10/6161—Compensation of chromatic dispersion

Definitions

  • the present application relates to optical communication technology, in particular to a method, device, receiver and storage medium for chromatic dispersion estimation.
  • IP Internet Protocol
  • IPTV Internet Protocol
  • mobile Internet large private network and other new applications from different fields are increasingly popular, and the transmission bandwidth demand of metropolitan area networks and backbone networks continues to grow .
  • Optical fiber communication technology with its advantages of ultra-high speed, large capacity, long distance, high anti-interference performance and low cost, has become the best choice to solve the capacity pressure of the backbone network.
  • optical fiber communication systems based on data coherent detection technology have become research hotspots in the field of optical communication.
  • the combination of digital coherent detection technology, multi-level signal modulation format, and polarization multiplexing (PM) technology can provide multiplied system communication capacity. Since 2010, the 100G optical communication module based on PM-QPSK (Quadrature phase shift keyin, QPSK for short) has been gradually commercialized, and research work on higher single-channel 400G and even IT has also become a hot spot.
  • PM-QPSK Quadrature phase shift keyin, QPSK for short
  • fiber dispersion refers to the phenomenon that different components in the optical pulse signal from the transmitting end are transmitted in the optical fiber at different rates and reach the receiving end at different times.
  • the dispersion effect makes the transmitted signal blurred in the receiving end after being transmitted in the optical fiber. This blurring causes inter-symbol interference, which in turn leads to power cost.
  • Dispersion is a cumulative effect. The longer the transmission link, the more significant the amount of dispersion.
  • the wide impulse response of the accumulated dispersion may be dispersed in hundreds or even thousands of symbols, which requires dispersion compensation.
  • accurate dispersion compensation can ensure reliable clock recovery and carrier synchronization, and is also very important for subsequent polarization equalization. Inappropriate dispersion value compensation may lead to complete digital coherent receivers. failure. Therefore, accurate dispersion value estimation is the first step in the normal operation of the system.
  • the embodiments of the present application provide a simple, fast, and high-precision dispersion estimation method, device, receiver, and storage medium.
  • a dispersion estimation method includes: obtaining frequency domain data, obtaining data to be processed according to the orthogonal two polarization state data contained in the frequency domain data; filtering the processed data, and converting the spectrum width between adjacent frequency points Broaden the setting multiple; extract the filtered data to be processed to obtain the target data set; calculate the target function corresponding to the set interval value according to the target data set, and calculate the dispersion estimation value according to the phase angle value corresponding to the target function.
  • a dispersion value estimation device includes: a frequency domain data acquisition module for acquiring frequency domain data, and obtaining data to be processed according to two orthogonal polarization data contained in the frequency domain data; a filtering module for performing the above processing The data is filtered, and the spectrum width between adjacent frequency points is expanded to set a multiple; the extraction module is used to extract the filtered data to be processed to obtain the target data group; the dispersion estimation module is used to calculate based on the target data group The target functions corresponding to the set interval values are respectively calculated according to the phase angle values corresponding to the above target functions.
  • An optical communication data coherent receiver includes an IQ imbalance compensation module, a dispersion estimation and compensation module, a polarization demultiplexing module, a carrier recovery module and a decision decoding module connected in sequence, wherein the above dispersion estimation and compensation
  • the module is used to implement the above dispersion estimation method in the embodiment of the present application.
  • a storage medium stores executable instructions in the storage medium, and when the executable instructions are executed by a processor, the foregoing dispersion estimation method according to an embodiment of the present application is implemented.
  • the dispersion estimation method, device, optical communication data coherent receiver and storage medium obtained in the above embodiments obtain the data to be processed according to the orthogonal two polarization state data contained in the frequency domain data by acquiring the frequency domain data. Filter the to-be-processed data, widen the spectrum width between adjacent frequency points to set a multiple, extract the filtered to-be-processed data to obtain a target data group, and calculate the target functions corresponding to the set interval values according to the target data group Calculate the dispersion estimate based on the phase angle value corresponding to the above objective function. In this way, by processing and filtering the data in the frequency domain, the spectrum width of the frequency point is widened, which can avoid the problem that the system is sensitive to the sampling spectrum when processing in the frequency domain.
  • FIG. 1 is a schematic structural diagram of an optical communication data coherent receiver in an embodiment of the present application.
  • FIG. 6 is a flowchart of a dispersion estimation method in yet another optional specific embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of a dispersion estimation device in an embodiment of the present application.
  • the adaptive dispersion estimation algorithm provided in the related art mainly includes: first, based on Dispersion estimation method of dispersion scanning; second, data-assisted dispersion estimation algorithm; third, and adaptive dispersion compensation algorithm based on Godard Clock (Tone, GCT) best matching.
  • the dispersion based on dispersion scanning it is necessary to try different dispersion values and monitor the objective function, which takes a long time, and the progress will be limited by the number of attempts, and the accuracy cannot be guaranteed, so the dispersion based on dispersion scanning
  • the estimation method has the problems of long dispersion value estimation time and low dispersion estimation accuracy.
  • the data-based dispersion estimation algorithm needs to rely on the training sequence, that is, it needs to use the prior information of the channel to calculate the channel dispersion. Response, and then solve the solution of the filter coefficients in the time or frequency domain.
  • the periodic signal has a strong spectral component at its clock frequency called GCT.
  • GCT the amplitude or power of the GCT (the square of the amplitude) is extremely sensitive to the residual dispersion in the signal. It can be used as a dispersion estimation parameter.
  • the dispersion estimation method based on the best matching of the GCT is sensitive to the sampling frequency deviation in high-speed systems. Whether the GCT best matching dispersion estimation method uses dispersion scanning or direct calculation, the complexity is higher.
  • the wide impulse response of long-distance transmission of cumulative chromatic dispersion may be scattered in hundreds or even thousands of symbols, and is proportional to the square of the signal sampling frequency.
  • the frequency domain compensation is fast Fourier transform (Fast Fourier Transformation, FFT for short) is very long, because the length of FFT is large and the frequency spectrum is dense, so that the system is sensitive to the sampling spectrum when processing in the frequency domain.
  • FFT Fast Fourier Transformation
  • the inventor of the present application optionally found that by processing and filtering data in the frequency domain to broaden the spectrum width of the frequency point, the problem of sensitivity to the sampling spectrum in the dispersion estimation method can be solved, thereby providing a Simple and fast, high-precision dispersion estimation method, and dispersion estimation device, optical communication data coherent receiver and storage medium for implementing the dispersion estimation method.
  • an embodiment of the present application provides an optical communication data coherent receiver, including an IQ imbalance compensation module 11, a dispersion estimation and compensation module 12, a polarization demultiplexing module 13, and a carrier recovery module 14 connected in sequence ⁇ Decision decoding module 15.
  • the IQ imbalance compensation module 11 is used to compensate IQ imbalance.
  • ADC analog-to-digital converter
  • the baseband signal is converted into a digital signal through an analog-to-digital converter (ADC).
  • ADC analog-to-digital converter
  • the baseband signal is converted into a digital signal through an analog-to-digital converter (ADC).
  • ADC analog-to-digital converter
  • the baseband signal when down-converted to baseband, it will be separated into an in-phase signal and a quadrature signal.
  • This is a local oscillator (Local Oscillator) using two sine waves with the same gain and a phase difference of 90 degrees. Select the site as cosine, sine wave) reached.
  • these processes are performed in the field of simulation, errors will occur.
  • the cosine and sine waves used for down conversion will generate gain and phase errors between each other. At this time, these errors may have a serious impact on the performance of the receiver. This is called IQ imbalance.
  • the dispersion estimation and compensation module 12 is used to estimate the dispersion value using the dispersion-compensated data stream in the receiver system, and use the dispersion estimation value for the system's dispersion compensation.
  • the polarization demultiplexing module 13 is used to implement the polarization demultiplexing technology.
  • Polarization multiplexing refers to the fact that two beams of the same or different wavelengths can be transmitted independently in one fiber at the same time, thereby doubling the information transmission capacity of the fiber without adding additional bandwidth resources.
  • the multiplexing technology refers to combining multiple signals in different ways or distinguishing methods at the signal sending end, and after undergoing the same channel transmission, separating the originally multiplexed signal at the receiving end to achieve the purpose of effective utilization.
  • the demultiplexing method mainly includes direct detection and coherent detection.
  • the carrier recovery module 14 is used for carrier recovery to recover the original carrier from the modulated signal.
  • carrier recovery There are two main methods of carrier recovery, one is to send a digital signal sequence at the same time as the carrier or the pilot signal related to it, and a narrowband filter or phase-locked loop can be used to directly extract the carrier at the receiver; one The received signal is a modulated signal that suppresses the carrier wave, and the coherent carrier wave is obtained by performing a non-linear transformation on the digital signal or using a special phase-locked loop.
  • the decision decoding module 15 is used to achieve the optimal or near optimal decoding of the error correction code.
  • Decision decoding mainly includes soft decision decoding and hard decision decoding. Among them, soft decision decoding refers to the use of digital data to decode the error correction code, and hard decision decoding refers to the decoder using the code's algebraic structure. Decode the error correction code.
  • an embodiment of the present application provides a dispersion value estimation method, which can be applied to the receiver shown in FIG. 1.
  • the dispersion value estimation method provided by the embodiment of the application can be applied to a receiver
  • the dispersion value estimation method includes the following steps:
  • Step 101 Obtain frequency domain data, and obtain data to be processed according to orthogonal two polarization state data contained in the frequency domain data;
  • the frequency domain data may be frequency domain data that is already used in dispersion compensation, or frequency domain data obtained by performing frequency domain conversion on time domain data.
  • the two orthogonal polarization data are independent of each other.
  • the receiver obtains the frequency domain data, and obtains the data to be processed according to the orthogonal two polarization state data contained in the frequency domain data.
  • the receiver obtains the frequency domain data and preprocesses the frequency domain data.
  • the two polarization state data are separated as the data to be processed; or, the two independent polarization data are linearly combined through preprocessing to form a multi-channel linear combination of the two polarization state components to be processed data.
  • the time domain data is converted into the frequency domain to obtain the frequency domain data can be achieved by Fast Fourier Transform (Fast Fourier Transformation, referred to as FFT), the two polarization state data can be represented by X[K] and Y[K] .
  • FFT Fast Fourier Transform
  • Step 103 Filter the data to be processed, and expand the spectrum width between adjacent frequency points to set a multiple;
  • the purpose of filtering the data to be processed is to broaden the spectrum range corresponding to the frequency point to increase the ability to resist sampling frequency deviation.
  • the setting multiple that broadens the spectrum width between adjacent frequency points can be achieved by configuring filter coefficients, which can be calculated or simulated in advance, or can be preset according to empirical values, such as the setting multiple can be 2 times or 3 times.
  • the data to be processed obtained according to step 101 usually includes two or more groups. When subsequent processing such as filtering is performed on the data to be processed, the pointer is independently processed for each group of the data to be processed.
  • Step 105 Extract the filtered data to be processed to obtain a target data group
  • extracting the filtered data to be processed means dividing the filtered output data into multiple groups, and extracting one of the multiple groups.
  • the to-be-processed data output after filtering can be extracted according to the setting of the filter coefficient in the data filtering.
  • the data to be processed obtained according to step 101 usually includes two or more groups, and the data to be processed is subjected to subsequent processing.
  • the pointer performs independently on each group of data to be processed, and the filtered data is processed accordingly.
  • Processing data for extraction is also a pointer for each group of data to be processed independently.
  • Step 107 Calculate the target function corresponding to the set interval value according to the target data set, and calculate the dispersion estimate according to the phase angle value corresponding to the target function.
  • the set interval value may include a plurality of, the target function includes a plurality of target functions corresponding to the set interval value respectively, the receiver calculates the target function corresponding to the set interval value according to the target data set, and calculates the dispersion according to the phase angle value corresponding to the target function
  • the estimated value includes: the receiver calculates the phase angle value of the target function separately, and merges the phase angle values of the target functions corresponding to the multiple set interval values respectively to obtain the phase angle value corresponding to the final target function to calculate the dispersion estimate value.
  • the dispersion estimation method obtains the data to be processed according to the orthogonal two polarization state data contained in the frequency domain data by acquiring the frequency domain data, filtering the data to be processed, and dividing the spectrum width between adjacent frequency points Broaden the setting multiple, extract the filtered data to obtain the target data group, calculate the target function corresponding to the set interval value according to the target data group, and calculate the dispersion estimate according to the phase angle value corresponding to the target function.
  • Processing and filtering the data in the frequency domain to widen the spectrum width of the frequency point can avoid the problem of the system being sensitive to the sampling spectrum when processing in the frequency domain, achieve accurate dispersion estimation, and calculate the dispersion estimation by the phase angle value corresponding to the objective function Value, which simplifies the calculation amount of the dispersion estimation algorithm, realizes the accurate estimation of the dispersion value in a simpler method, saves system resources, and reduces system power consumption.
  • obtaining data to be processed according to the orthogonal two polarization state data contained in the frequency domain data includes:
  • the X-polarization data and the Y-polarization data in the frequency domain data are separately divided into odd and even according to the index, the frequency domain data of the even index is inverted and the frequency domain data of the odd index is unchanged, the pre-processed X polarization state is obtained Data and Y polarization data;
  • the pre-processed X-polarized data and Y-polarized data are respectively used as data to be processed; or, the pre-processed X-polarized data and Y-polarized data are linearly combined to obtain multiple sets of data to be processed.
  • pre-processing the frequency-domain data to obtain the data to be processed includes dividing the X-polarization data and the Y-polarization data in the acquired frequency domain data by odd and even according to the indexes, and multiplying all the even-indexed frequency domain data by -1 and the odd-indexed frequency domain data is unchanged, and the pre-processed X polarization data and Y polarization data are obtained.
  • N is the frequency domain data length.
  • the two polarization states are independent of each other, and the pre-processed X-polarized state data X[K] and Y-polarized state data Y[K] are used as two sets of data to be processed; alternatively, the pre-processed X-polarized state data X[K] K] and Y polarization data Y[K] are linearly combined to obtain multiple sets of data to be processed in multiple different directions.
  • the pre-processed X polarization data and Y polarization data are linearly combined to obtain multiple sets of data to be processed, including:
  • the inverted data of the X polarization state and the data of the Y polarization state are used as the real part and the imaginary part of the complex number to obtain the third data to be processed.
  • the linear combination operation mainly includes inverse, linear addition and complex addition of corresponding polarization state data, thereby obtaining three sets of pending data containing X polarization state data and/or Y polarization state data in three different directions data.
  • the first to-be-processed data X 1 [K] is X polarization state data, or Y polarization state data is obtained by inversion.
  • X polarization state data as an example, the following public data can be obtained:
  • the second to-be-processed data X 2 [K] can be obtained by linearly adding the X polarization state data and the Y polarization state data, and can be obtained by using the following formula:
  • the third to-be-processed data X 3 [K] can be obtained by adding the inverted data of the X polarization state data and the Y polarization state to a complex number, and can be obtained by the following formula:
  • X polarization data and Y polarization data can be interchanged, such as the first to be processed
  • the data may be Y polarization state data Y[K]
  • the third data to be processed may invert the Y polarization state data as the real part of the complex number, and use the X polarization state data as the imaginary part of the complex number.
  • the dispersion value estimation is performed respectively, which can avoid the influence of polarization dispersion on the chromatic dispersion estimation.
  • filtering the data to be processed and setting a multiple of the spectrum width between adjacent frequency points includes:
  • the frequency domain filter coefficient is determined according to the frequency domain convolution corresponding to the time domain data windowing function processing corresponding to the frequency domain data, and the data to be processed is filtered according to the frequency domain filter coefficient.
  • the frequency domain filter coefficients are determined according to the frequency domain convolution corresponding to the time domain data corresponding to the frequency domain data plus the window function processing of a specified length.
  • the window is determined according to the data length of the frequency domain and the setting multiple of the pre-spreading the spectral width between adjacent frequency points.
  • the function includes determining the length of the window function according to the ratio N/M of the frequency domain data length and the set multiple of the spectrum width between adjacent frequency points, and the type of the window function can be selected according to the specific situation, such as rectangular Windows, Hamming windows, etc.
  • the width of the frequency domain between adjacent frequency points in the frequency domain is expanded to M times the original, then the filter coefficient is the frequency corresponding to the time domain data corresponding to the frequency domain data plus a window function of N/M length Domain filter coefficients, that is, the effect of filtering in the frequency domain is equivalent to adding a window function to the data in the middle of the data in the time domain, so as to achieve the frequency range corresponding to the frequency point to be expanded to the original M times to increase
  • the ability to resist sampling frequency offset, in the subsequent dispersion estimation, to avoid the increase of the target function error or the total error due to the existence of sampling frequency offset, the conjugate multiplication of the data and the frequency offset, by adding two frequency points The width of the spectrum can solve the problem of frequency offset.
  • step 107 the target functions corresponding to the set interval values are calculated according to the target data set, including:
  • Step 1071 autocorrelating the target data set
  • Step 1072 Multiply and accumulate the autocorrelation target data set according to the set interval value to obtain the target function corresponding to the set interval value.
  • Autocorrelation refers to the dependence between the instantaneous value of a signal at one moment and the instantaneous value of another moment.
  • the size of the setting interval value can be determined according to the actual application. Generally, the smaller the setting interval value, the larger the estimated value range and the correspondingly lower accuracy; otherwise, the larger the setting interval value, the smaller the estimated value range and the corresponding accuracy Higher.
  • the number of the setting interval values may be multiple, so as to obtain target functions corresponding to the multiple setting interval values, respectively.
  • the set interval value is represented by ⁇ 1.
  • the target data group after autocorrelation is set according to the set interval value Multiply and accumulate the conjugate to get the formula (6) of the objective function corresponding to the set interval value as follows:
  • Rx[n] represents the data after autocorrelation
  • the set interval value includes multiple groups, such as ⁇ 2 , ⁇ 3 ,...
  • the corresponding objective functions F 2 , F 3 ,... can be obtained, where the calculation formula is the same as the above formula (6), and will not be repeated here.
  • the autocorrelation output is conjugate multiplied according to the set interval value and accumulated to obtain the target function corresponding to the set interval value, so that the size and number of the set interval value can be determined according to the actual application to adjust the estimated value The accuracy of the range.
  • step 1071, autocorrelating the target data set includes: conjugate multiplying the data of the interval baud rate in the frequency domain in the target data set.
  • the target data group refers to the data obtained after extracting the data to be processed.
  • Autocorrelation of the target data set means that the data of the interval baud rate in the frequency domain is conjugated and multiplied.
  • the data to be processed X[K], Y[K ] The target data set obtained by filtering and extraction are represented by DX[L] and DY[L] respectively.
  • X polarization data X′[K] as an example, the autocorrelation of the target data set DX[L] can be obtained as follows Formula (7) realizes:
  • conj( ⁇ ) means conjugate the data.
  • the setting interval value includes a plurality of different setting interval values
  • calculating the dispersion estimation value according to the phase angle value corresponding to the objective function includes:
  • Step 1073 according to whether the relationship between the phase angle values of the target functions corresponding to the different setting interval values meets the setting conditions, and merge to obtain the target phase angle value of the target function;
  • the dispersion estimate value is calculated according to the target phase angle value, the set interval value corresponding to the target phase angle value, the speed of light, the center wavelength of the optical signal, the sampling frequency, the discrete Fourier transform length, and the symbol rate.
  • the setting interval value is a plurality of setting interval values of different sizes.
  • the corresponding coefficient is determined by the setting interval value corresponding to the target phase angle value, the speed of light, the center wavelength of the optical signal, the sampling frequency, the discrete Fourier transform length and the symbol rate.
  • the target phase angle value and the target phase angle value Set the interval value, speed of light, center wavelength of the optical signal, sampling frequency, discrete Fourier transform length and symbol rate.
  • Equation 8 The formula for calculating the dispersion estimate value is shown in Equation 8 below:
  • c is the speed of light and the unit is meter/second (m/s); ⁇ is the center wavelength of the optical signal corresponding to the frequency domain data and the unit is nanometer (nm); nfft is the discrete Fourier transform length; f s is the sampling Frequency, unit is GHz, f d represents symbol rate, unit is Gbaud; the unit of the final output dispersion estimate CD is nanosecond/nanometer (ns/nm).
  • step 1073 according to whether the relationship between the phase angle values of the target functions corresponding to the different setting interval values meets the setting conditions, the combination is performed to obtain the target phase angle value of the target function, including:
  • phase angle value of the target function corresponding to the two adjacent setting interval values is used as the initial first phase angle value and the second phase angle value;
  • the second phase angle value and the integer part are combined and subtracted by 1 as the updated second phase angle value
  • the second phase angle value and the integer part are combined and added by 1 as the updated second phase angle value
  • the second phase angle value and the integer part are combined as the updated second phase angle value
  • the number of set interval values is w (w>1), and ⁇ 1 ⁇ 2 ⁇ ... ⁇ w .
  • the phase angle value of the objective function F1 corresponding to the set interval value ⁇ 1 is Set the phase angle value of the objective function F2 corresponding to the interval value ⁇ 2
  • set the phase angle value of the objective function F3 corresponding to the interval value ⁇ w is
  • angle( ⁇ ) represents the phase angle of the data
  • floor( ⁇ ) represents the rounding down
  • itr represents the integer obtained by the product of the ratio of the two set interval values and the first phase angle value Part and u represent the fractional part
  • n 2, 3,...w.
  • step 1074 before calculating the dispersion estimation value according to the target phase angle value, the set interval value corresponding to the target phase angle value, the speed of light, the center wavelength of the optical signal, the sampling frequency, the discrete Fourier transform length and the symbol rate ,include:
  • the target phase angle value obtained the previous time is used as the final target phase angle value.
  • the objective phase angle value of the most recently calculated objective function may refer to the objective phase angle value of the NL objective functions calculated most recently through the current time, where the value of NL may be set according to actual needs .
  • the currently calculated target phase angle value is selected as the final target phase angle value, so that the corresponding value is obtained according to the current calculation Calculate the estimated dispersion value of the target phase angle value;
  • the previous target phase angle value is used as the final target phase angle value Value, and accordingly calculate the dispersion estimation value based on the target phase angle value obtained last time.
  • the dispersion estimation value before calculating the dispersion estimation value according to the phase angle value corresponding to the objective function, it includes:
  • a low-pass filter coefficient is used as the filter coefficient for smooth filtering the objective function.
  • the filter coefficient for smooth filtering the objective function.
  • the objective function as F 1 , F 2 , F 3 ,...F w for example
  • the filtered output obtained after smoothing filtering is F 1 ', F 2 ', F 3 ',...F w '.
  • the dispersion estimation value before calculating the dispersion estimation value according to the phase angle value corresponding to the objective function, it includes:
  • the set length of the register is the same as the number of calculations of the phase angle value of the objective function stored in the register for the most recent calculation.
  • the number of registers corresponds to the number of objective functions. Taking the objective function as F 1 , F 2 , F 3 ,...F w , and storing the phase angle value of the objective function calculated in the most recent NL times into the register as an example, construct W registers of length NL and use buffer1 ,Buffer2,...,buffer W means to store the objective functions F 1 ,F 2 ,F 3 ,...F w calculated from the most recent NL times.
  • the initial values in buffer1-W are all NL zeros, and the objective functions corresponding to the accumulation result in buffer1-W are F 1_sum , F 2_sum , F 3_sum ,...F w_sum , and then obtained according to the accumulation result in buffer1-W
  • the objective functions F 1_sum , F 2_sum , F 3_sum ,...F w_sum calculate the objective phase angle values of the objective function.
  • comparing the target phase angle values of the most recently calculated objective functions may be the objective functions F 1_sum , F 2_sum , F 3_sum corresponding to the accumulation result in buffer1-W, ...F w_sum 's phase angle value is calculated to compare the most recent NL target phase angle values.
  • the current Calculated target phase angle value As the final target phase angle value; when the absolute value of the difference between any two of the NL target phase angle values is greater than the threshold, the previous target phase angle value is obtained As the final objective phase angle value.
  • the dispersion estimation value after calculating the dispersion estimation value according to the phase angle value corresponding to the objective function, it includes:
  • the dispersion estimation value obtained the previous time is maintained.
  • the objective phase angle value of the most recently calculated objective function may refer to the objective phase angle value of the NL objective functions calculated most recently through the current time, where the value of NL may be set according to actual needs .
  • the dispersion estimate value is updated; otherwise, the update of the dispersion estimate value is stopped and the previous dispersion estimate value is maintained.
  • step 105 the filtered data to be processed is extracted to obtain a target data set, including:
  • the filtered data to be processed is divided into the same number of data groups as the set multiple, and the target data group is extracted from the data group.
  • data extraction is related to the setting of filter coefficients in data filtering. If the spectrum width between two frequency points in the frequency domain is changed to the original M times, the filtered data to be processed is divided into M groups, and the data at intervals M is used as a group.
  • the data group obtained by dividing the data to be processed into M groups can be expressed as follows:
  • the index of L ⁇ M+1 is regarded as the first group
  • the index of L ⁇ M+2 is regarded as the second group
  • the index of L ⁇ M+M is regarded as the Mth group.
  • the target data group can be extracted from the data group by selecting the Pth group in the M group as the target data group.
  • P can be pre-configured and set to any value from 1 to M.
  • the dispersion estimation method includes the following steps :
  • Step S11 data preprocessing, using X polarization data and Y polarization data as data to be processed respectively;
  • Obtain the frequency domain data and divide the acquired frequency domain data into odd and even according to the index through data preprocessing. Multiply all the even index frequency domain data by -1, and the odd index frequency domain data will not change.
  • the data in high-speed ultra-high-speed optical fiber communication is divided into two intersecting polarization states, X polarization state and Y polarization state, and the two polarization states are independent of each other. Taking X polarization data as an example, the operation of Y polarization data is the same.
  • Step S12 Filter and output the to-be-processed data output after data preprocessing
  • FIR filtering is performed separately on the pre-processed X polarization state data X[K] and Y polarization state data Y[K].
  • the purpose of filtering is to widen the frequency range corresponding to the frequency point, thereby increasing the ability to resist sampling frequency deviation.
  • the data in the frequency domain needs to be conjugated and multiplied by the interval baud rate data in the frequency domain. Because of the sampling frequency offset, the time-frequency point shift of the conjugate multiplication may be caused, which will increase the error of the objective function or be completely wrong. By increasing the spectrum width between two frequency points, the problem of frequency offset can be avoided.
  • the filter coefficient can be configured according to the specific conditions of the system.
  • the frequency point needs to be expanded to several times the original can be calculated or simulated in advance, this value can also be generally 2 times or 3 times according to experience.
  • the filter coefficient is the corresponding frequency domain filter coefficient after the time domain plus the N/M length window, that is, the effect after the frequency domain filtering is equal
  • a window function is added to the data in the middle N/M of the data in the time domain.
  • the window function can be selected according to specific circumstances, such as a rectangular window, a Hamming window, and so on.
  • N is the data length in the frequency domain
  • M is an integer greater than or equal to 2.
  • Step S13 Extract the data output by the data filtering
  • data extraction is performed on the data filtering output.
  • Data extraction is performed separately for X polarization data and Y polarization data.
  • the data output by the data filtering is divided into M groups, and the data with the interval M is taken as a group.
  • the index of L ⁇ M+1 is taken as the first group
  • the index of L ⁇ M+2 is taken as the second group
  • the index of L ⁇ M+M is taken as the Mth group.
  • Select the Pth group among the M groups as the output, and P can be set to any value from 1 to M. It should be noted that once the P group is selected as the output, all subsequent data extraction operations must select the P group as the output. P as a parameter can only be configured when the system is started or restarted, and other conditions cannot be changed.
  • Step S14 performing an auto-correlation operation on the extracted data
  • conj( ⁇ ) means conjugate the data.
  • Step S15 multiply and accumulate the data after performing the autocorrelation operation according to the set interval value to obtain the objective function
  • the objective function is calculated according to the interval value ⁇ 1 set in advance:
  • n 2,3,...,w, where angle( ⁇ ) represents the phase angle of the data and floor( ⁇ ) represents the downward rounding.
  • c is the speed of light and the unit is meter/second (m/s); ⁇ is the center wavelength of the optical signal corresponding to the frequency domain data and the unit is nanometer (nm); nfft is the discrete Fourier transform length and f s is the sampling Frequency, unit is GHz, f d represents symbol rate, unit is Gbaud; the unit of final output dispersion value CD is nanoseconds/nanometer (ns/nm).
  • the dispersion estimation method includes the following steps:
  • Step S21 data preprocessing, linearly combining X polarization data and Y polarization data to obtain three sets of data to be processed;
  • Obtain the frequency domain data and divide the acquired frequency domain data into odd and even according to the index through data preprocessing. Multiply all the even index frequency domain data by -1, and the odd index frequency domain data will not change.
  • the data in high-speed ultra-high-speed optical fiber communication is divided into two intersecting polarization states, X polarization state and Y polarization state, and the two polarization states are independent of each other. Taking X-polarized data as an example, multiply the even-indexed frequency domain data in X-polarized data by -1 and the odd-indexed frequency domain data unchanged to obtain pre-processed X-polarized data.
  • the pre-processed X polarization state data and Y polarization state data are linearly combined as follows:
  • Step S22 filter and output the data to be processed after the data preprocessing; the difference from step S12 is that the filtered input data is replaced with X 1 [k], X 2 [k], X 3 [k], Correspondingly, the output result of data filtering also becomes three groups accordingly.
  • Step S23 extract the data output by the data filtering; the difference from step S13 is that the input data is replaced with the three sets of output results of step S22, and the extracted output results also become three sets;
  • Step S24 perform an autocorrelation operation on the extracted data; the difference from step S14 is that the data input from the autocorrelation operation is replaced with the three sets of extraction output results of step S23;
  • Step S25 the data after the autocorrelation operation is conjugated and multiplied according to the set interval value and accumulated to obtain the target function; the difference from step S15 is that the target function is calculated separately according to each set of inputs to obtain three sets of target functions;
  • F′ 1 F 1_1 +F 2_1 +F 3_1
  • F′ 2 F 2_1 +F 2_1 +F 2_1
  • F′ 3 F 3_1 +F 3_1 +F 3_1
  • F′ 1 , F′ 2 , F′ 3 ,...F′ w are output as the objective function.
  • Step S26 Perform phase calculation according to the phase angle value corresponding to the objective function to obtain the final objective function phase angle value; the difference from step S16 is that the input of the phase calculation is replaced with the objective function F′ 1 ,F′ output from step S25 2 ,F′ 3 ,...F′ w .
  • Step S27 calculate and output the dispersion estimation value according to the target function phase angle value output by the phase calculation; the same as step S17.
  • the dispersion estimation method includes the following steps:
  • Step S31 data preprocessing, using X polarization data and Y polarization data as data to be processed respectively; the same as step S11.
  • Step S32 filtering and outputting the data to be processed output after data preprocessing; the same as step S12.
  • Step S33 extract the data output by the data filtering; the same as step S13.
  • Step S34 performing an auto-correlation operation on the extracted data; the same as step S14.
  • Step S35 the data after the autocorrelation operation is conjugated, multiplied and accumulated according to the set interval value, and the target function is obtained, then smoothed and filtered, and the filtered target function is output;
  • the difference from step S15 lies in:
  • the interval value such as ⁇ 2 , ⁇ 3 ,... ⁇ w calculated objective function F 2 ,F 3 ,...F w , smooth filtering, filter coefficients using low-pass filter coefficients, the filter output is F ′ 1 ,F′ 2 ,F′ 3 ,...F′ w ;
  • Step S36 Perform phase calculation according to the phase angle value corresponding to the filtered target function to obtain the final phase angle value of the target function; the difference from step S16 is that the input of the phase calculation is replaced with the target function F′ 1 output in step S35 ,F′ 2 ,F′ 3 ,...F′ w .
  • Step S37 Calculate the dispersion estimation value according to the phase angle value of the objective function output from the phase calculation, and update the dispersion estimation value according to the relationship between the absolute value of the difference between any two of the recent multiple phase angle values and the threshold
  • the difference from step S17 is that after calculating the dispersion estimation value according to the phase angle value of the objective function output by the phase calculation, the most recent phase angle values Compare if multiple phase angle values If the absolute value of the difference between any two is less than the threshold Th, the dispersion estimate value will be updated, and the currently calculated dispersion estimate value will be output; if there are multiple phase angle values If the absolute value of the difference between any two is greater than or equal to the threshold value Th, the update of the dispersion estimate value is stopped, the previous dispersion estimate value is maintained, and the last dispersion estimate value is output.
  • the dispersion estimation method includes the following steps:
  • Step S41 data preprocessing, using X polarization data and Y polarization data as data to be processed respectively; the same as step S11.
  • Step S42 filtering and outputting the data to be processed output after data preprocessing; the same as step S12.
  • Step S43 extract the data output by the data filtering; the same as step S13.
  • Step S44 perform an auto-correlation operation on the extracted data; the same as step S14.
  • Step S45 multiply and accumulate the data after the autocorrelation operation according to the set interval value to obtain the objective function; the same as step S15.
  • Step S46 construct a plurality of registers of length NL to store the latest NL calculated target function, accumulate the accumulated target function according to the data in the register, and calculate the phase according to the phase angle value corresponding to the accumulated target function to obtain The final phase angle value of the objective function; the difference from step S16 is:
  • the initial values in buffer1-W are all NL 0s.
  • the accumulated objective function according to the data in the register is to sum the data in buffer1-W to get F 1_sum , F 2_sum , F 3_sum ,...F w_sum .
  • phase calculation according to the phase angle value corresponding to the accumulated objective function means that the objective functions F 1_sum , F 2_sum , F 3_sum ,...F w_sum are respectively calculated according to the objective functions F 1_sum , F 2_sum , F 3_sum ,...F w_sum value
  • the calculation process is the same as step S16.
  • Step S461 the relationship between the absolute value of the difference between any two of the phase angle values of the multiple target functions calculated according to the data in the register and the threshold value to update the final phase angle value of the target function; different from step S16 Lies in:
  • the method further includes: multiple objective function phase angle values calculated according to the data in buffer1-W When the absolute value of the difference between any two is less than the threshold Th, then the As a phase calculation output; otherwise, use the last As a phase calculation output.
  • Step S47 calculate and output the dispersion estimation value according to the phase angle value of the target function output by the phase calculation; the difference from step S17 is that: replace the target function output by the phase calculation with the target function phase value output according to step S461, where the dispersion estimate value Is calculated in the same way as step S17.
  • a wide impulse response that transmits accumulated dispersion over a long distance may be dispersed in hundreds or even thousands of symbols, and is proportional to the square of the signal sampling frequency.
  • the FFT length when implementing frequency domain compensation will be large, so the frequency spectrum is dense, so that the system is sensitive to the sampling spectrum when processing in the frequency domain, so the traditional dispersion estimation method cannot be used in ultra-high-speed optical fiber communication systems. normal work.
  • the dispersion estimation method provided by the above embodiment of the present application solves the problem that the receiver performs data preprocessing and filtering in the frequency domain to widen the frequency spectrum width of the frequency point, thereby solving the sensitivity of the ultra-high-speed long-distance optical fiber system dispersion estimation method to sampling frequency deviation
  • the problem of realizing accurate dispersion estimation of ultra-high-speed optical fiber communication is calculated.
  • the calculation amount of the dispersion estimation algorithm is simplified, and the accurate estimation of the dispersion value of the ultra-high-speed long-distance optical fiber communication system is realized in a simpler method, saving System resources reduce system power consumption.
  • Table 1 it is a comparison table of the system resources required to calculate the primary dispersion estimation value using the dispersion estimation method provided in the embodiments of the present application and the dispersion estimation method based on FFT operation in the related art, taking the data bit width as an example
  • the amount of calculation for calculating the primary dispersion estimation value according to the target function value is about 78.6% less than that for calculating the primary dispersion estimation value based on the traditional FFT method.
  • a chromatic dispersion estimation device which includes a frequency domain data acquisition module 121 for acquiring frequency domain data according to the orthogonal two polarization state data contained in the frequency domain data Obtain the data to be processed; the filtering module 123 is used to filter the data to be processed, and the spectrum width between adjacent frequency points is expanded to set a multiple; the extraction module 125 is used to extract the filtered data to be processed to obtain the target data group Dispersion estimation module 127, used to calculate the target function corresponding to the set interval value according to the target data set, and calculate the dispersion estimation value according to the phase angle value corresponding to the target function.
  • the frequency domain data acquisition module 121 includes a polarization data unit, a first preprocessing unit, or a second preprocessing unit.
  • the polarization data unit is used to convert the X polarization data and the Y polarization in the frequency domain data.
  • the state data is divided into odd and even according to the index respectively, the frequency domain data of the even index is inverted and the frequency domain data of the odd index is unchanged, the pre-processed X polarization data and Y polarization data are obtained; the first preprocessing unit , Used to pre-process the X-polarized data and Y-polarized data respectively as data to be processed; the second pre-processing unit is used to perform linear combination operation on the pre-processed X-polarized data and Y-polarized data, Get multiple sets of data to be processed.
  • the second pre-processing unit is specifically configured to use the X polarization state data or the Y polarization state data as the first data to be processed; the X polarization state data and the Y polarization state data are inverted and added to obtain the second To-be-processed data; the X-polarized state data and the Y-polarized state data are inverted and used as the real part and imaginary part of the complex number to obtain the third to-be-processed data.
  • the filtering module 123 is specifically configured to determine the window function according to the frequency domain data length and the setting multiple of pre-expanding the spectral width between adjacent frequency points; according to the time domain data corresponding to the frequency domain data, the window function
  • the frequency domain convolution corresponding to the processing determines the frequency domain filter coefficients, and the data to be processed is filtered according to the frequency domain filter coefficients.
  • the dispersion estimation module 127 includes an autocorrelation unit 1271 and an objective function acquisition unit 1272.
  • the autocorrelation unit 1271 is used to autocorrelate the target data set;
  • the objective function acquisition unit 1272 is used to convert the autocorrelation
  • the target data set is conjugate multiplied and accumulated according to the set interval value to obtain the target function corresponding to the set interval value.
  • the autocorrelation unit 1271 is specifically used to conjugate and multiply the data of the interval baud rate in the frequency domain in the target data set.
  • the setting interval value includes a plurality of different setting interval values
  • the dispersion estimation module 127 further includes a phase calculation unit 1273 and a dispersion estimation unit 1274, and a phase calculation unit 1273, which is specifically configured to set the interval value differently according to Whether the relationship between the corresponding phase angle values of the objective function meets the setting conditions and merge to obtain the objective phase angle value of the objective function;
  • the dispersion estimation unit 1274 is used to set the interval value corresponding to the objective phase angle value and the objective phase angle value , The speed of light, the center wavelength of the optical signal, the sampling frequency, the discrete Fourier transform length and the symbol rate to calculate the dispersion estimate.
  • the phase calculation unit 1273 is specifically configured to use the phase angle value of the target function corresponding to the two adjacent setting interval values as the initial first phase angle value and the second phase angle value; according to the two settings
  • the product of the ratio of the interval value and the first phase angle value gives the corresponding integer part and fractional part; when the difference between the second phase angle value and the fractional part is greater than or equal to the first preset value, the second phase angle value and After the integer part is merged, minus 1 is used as the updated second phase angle value; when the difference between the second phase angle value and the decimal part is less than the second preset value, the second phase angle value and the integer part are merged and added as 1
  • the updated second phase angle value when the difference between the second phase angle value and the fractional part is less than the first preset value and greater than or equal to the second preset value, the second phase angle value and the integer part are combined as an update After the second phase angle value; the updated second phase angle value and the phase angle value of the objective function corresponding to the next set interval value
  • the phase calculation unit 1273 is specifically used to compare the target phase angle value of the recently calculated target function multiple times, and determine that the absolute value of the difference between any two is less than the threshold, the current The target phase angle value is regarded as the final target phase angle value; when it is determined that the absolute value of the difference between any two is greater than the threshold value, the target phase angle value obtained the previous time is used as the final target phase angle value.
  • the objective function acquisition unit 1272 is further configured to perform smooth filtering on the objective function before calculating the dispersion estimation value according to the phase angle value corresponding to the objective function.
  • the phase calculation unit 1273 is further configured to construct a plurality of registers with a set length before calculating the dispersion estimation value according to the phase angle value corresponding to the target function, and to calculate the phase angle value of the target function calculated most recently. Correspondence is stored in the register, and the phase angle value corresponding to the objective function is obtained according to the accumulation result of each register.
  • the dispersion estimation unit 1274 is further used to compare the target phase angle values of the most recently calculated target function after calculating the dispersion estimate value according to the phase angle value corresponding to the target function to determine any two When the absolute value of the difference is less than the threshold, the dispersion estimate is updated according to the currently obtained dispersion estimate; when the absolute value of any two differences is greater than the threshold, the previous dispersion estimate is maintained.
  • the extraction module 125 is specifically configured to divide the filtered data to be processed into the same number of data groups as the set multiple, and extract the target data group from the data group.
  • the dispersion estimation device provided in the above embodiment is only exemplified by the division of the above program modules. In practical applications, the above steps can be allocated by different program modules according to needs, that is, the The internal structure of the device is divided into different program modules to complete all or part of the processing described above.
  • the dispersion estimation device and the dispersion estimation method embodiments provided in the above embodiments belong to the same concept. For the specific implementation process, see the method embodiments, and details are not described here.
  • An embodiment of the present application further provides a computer storage medium, for example, including a memory storing a computer program, and the computer program can be executed by a processor to complete the steps of the dispersion estimation method provided in any embodiment of the present application.
  • the computer storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disk, or CD-ROM; it may also be various devices including one or any combination of the above memories.
  • the frequency domain data is obtained, and the data to be processed is obtained according to the orthogonal two polarization state data contained in the frequency domain data.
  • Filter the above-mentioned data to be processed expand the spectrum width between adjacent frequency points to set a multiple, extract the filtered to-be-processed data to obtain a target data group, and calculate the target data group respectively corresponding to the set interval value according to the target data group
  • the objective function calculates the dispersion estimate based on the phase angle value corresponding to the above objective function.

Landscapes

  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Optical Communication System (AREA)

Abstract

本申请实施例公开一种色散估计方法、装置、接收机及存储介质，该方法包括获取频域数据，根据频域数据包含的正交的两个偏振态数据得到待处理数据；对待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；对滤波后的待处理数据进行抽取得到目标数据组；根据目标数据组计算分别与设置间隔值对应的目标函数，根据目标函数对应的相角值计算色散估计值。

Description

色散估计方法、装置、接收机及存储介质
本申请要求于2018年11月30日提交中国专利局、申请号为201811449784.0、发明名称“色散估计方法、装置、接收机及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及光通信技术，尤其涉及一种色散估计方法、装置、接收机及存储介质。
背景技术
近年来，随着社会信息化程度的不断提高，基于IP(Internet Protocol)的数据业务呈爆炸式的增长。尤其是端到端视频播放和下载，网络电视(Internet Protocol Television，简称IPTV)网络，移动互联网，大型专网等来自不同领域新型应用的日益流行，城域网和骨干网的传输带宽需求持续增长。光纤通信技术以其超高速、大容量、长距离、高抗干扰性能和低成本等优点，成为解决骨干网容量压力的最佳选择。
得益于数字信号处理技术(Digital signal processing，简称DSP)和高速数据采集技术以及相关光器件的长足进展，基于数据相干检测技术的光纤通信系统成为光通信领域的研究热点。数字相干检测技术和多级信号调制格式、以及偏振复用(Palarization multiplexing，简称PM)技术相结合能够成倍地提供系统的通信容量。2010年以来，基于PM-QPSK(Quadrature phase shift keyin，简称QPSK)的100G光通信模块逐渐商用化，更高的单通道400G甚至IT的研究工作也成为热点。
光纤传输系统中，尽可能降低光纤信道的传输误码率，提升传输质量，从而增加信号传输距离是系统设计的一个重要指标。影响传输距离的因素主要有损耗、非线性效应和色散。其中，光纤色散是指来自发射端的光脉冲信号中的 不同成分以不同的速率在光纤中传输，在不同时刻到达接收端的现象。色散效应使得发射的信号在光纤中传输后，在接收端变得模糊，这种模糊引起了码间干扰，并进而导致了功率代价。色散是一种可累积的效应，传输的链路越长，色散量就越显著。
在没有添加光域色散补偿模块的链路中，累积色散的宽脉冲响应可能分散在数百甚至上千个码元中，这就需要进行色散补偿。在相干接收机的算法处理中，精确的色散补偿能保证可靠的时钟恢复和载波同步，对于后续的偏振均衡也是十分重要的，而不适当的色散值补偿可能会导致整个数字相干接收机的完全失败。因此，进行精确的色散值估计是系统正常工作的第一步。在下一代基于光交换的全光网络中，实现光路的动态配置是发展趋势，即发送端与接收端之间的信号路径会动态变化，链路中的累积色散值会动态变化，缓慢的估计过程会明显延缓后续处理模块的初始化，动态快速的色散估计算法就显得尤为重要。综上所述，快速、精确地估计出光纤链路中累积的色散值在相干光通信中至关重要。
发明内容
为解决现有存在的技术问题，本申请实施例提供一种简单快速、高精度的色散估计方法、装置、接收机及存储介质。
为达到上述目的，本申请实施例的技术方案是这样实现的：
一种色散估计方法，包括：获取频域数据，根据上述频域数据包含的正交的两个偏振态数据得到待处理数据；对上述处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；对滤波后的上述待处理数据进行抽取得到目标数据组；根据上述目标数据组计算分别与设置间隔值对应的目标函数，根据上述目标函数对应的相角值计算色散估计值。
一种色散值估计装置，包括：频域数据获取模块，用于获取频域数据，根据上述频域数据包含的正交的两个偏振态数据得到待处理数据；滤波模块，用于对上述处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；抽取模块，用于对滤波后的上述待处理数据进行抽取得到目标数据组；色散估计模块，用于根据上述目标数据组计算分别与设置间隔值对应的目标函数，根据上述目标函数对应的相角值计算色散估计值。
一种光通信数据相干接收机，上述接收机包括依次连接的IQ不平衡补偿模块、色散估计与补偿模块、偏振解复用模块、载波恢复模块和判决译码模块，其中，上述色散估计与补偿模块用于实现本申请实施例上述的色散估计方法。
一种存储介质，上述存储介质中存储有可执行指令，上述可执行指令被处理器执行时实现本申请实施例上述的色散估计方法。
上述实施例所提供的色散估计方法、装置、光通信数据相干接收机及存储介质，通过获取频域数据，根据上述频域数据包含的正交的两个偏振态数据得到待处理数据，对上述待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数，对滤波后的上述待处理数据进行抽取得到目标数据组，根据上述目标数据组计算分别与设置间隔值对应的目标函数，根据上述目标函数对应的相角值计算色散估计值，如此，通过在频域上对数据进行处理和滤波，展宽频点的频谱宽度，可以避免系统在频域处理时对采样频谱敏感的问题，实现精确色散估计，且通过根据目标函数对应的相角值计算色散估计值，简化了色散估计算法的计算量，以较简单的方法实现了色散值的准确估计，节省系统资源，降低系统功耗。
附图说明
图1为本申请一实施例中光通信数据相干接收机的结构示意图；
图2为本申请一实施例中色散估计方法的流程图；
图3为本申请另一实施例中色散估计方法的流程图；
图4为本申请一可选的具体实施例中色散估计方法的流程图；
图5为本申请另一可选的具体实施例中色散估计方法的流程图；
图6为本申请又一可选的具体实施例中色散估计方法的流程图；
图7为本申请再一可选的具体实施例中色散估计方法的流程图；
图8为本申请一实施例中色散估计装置的结构示意图。
具体实施方式
以下结合说明书附图及具体实施例对本申请技术方案做可选地的详细阐述。除非另有定义，本文所使用的所有的技术和科学术语与属于本申请的技术领域 的技术人员通常理解的含义相同。本文中在本申请的说明书中所使用的术语只是为了描述具体的实施例的目的，不是旨在于限制本申请。本文所使用的术语“和/或”包括一个或多个相关的所列项目的任意的和所有的组合。
在以下的描述中，涉及到“一些实施例”的表述，其描述了所有可能实施例的子集，但是应当理解，“一些实施例”可以是所有可能实施例的相同子集或不同子集，并且可以在不冲突的情况下相互结合。
为了能够快速、精确地估计出光纤链路中累积的色散值(Chromatic Dispersion，简称CD)，本申请发明人在研究中发现，相关技术中提供的自适应色散估计算法主要包括：第一、基于色散扫描的色散估计方法；第二、基于数据辅助的色散估计算法；第三、以及基于戈达尔钟音(Godard Clock Tone，简称GCT)最佳匹配的自适应色散补偿算法。其中，针对第一、基于色散扫描的色散估计方法，需要尝试不同的色散值并监控目标函数，所需时间较长，并且进度会受尝试次数的限制，精确无法保证，从而基于色散扫描的色散估计方法存在色散值估计时间较长，色散估计精度较低的问题；针对第二、基于数据辅助的色散估计算法，需要依赖于训练序列，即需要利用信道的先验信息，计算出信道的色散响应，然后在时域或频域中求解滤波器系数的解，这两种算法估计速度均较慢，对后续的相位恢复、信道均衡等模块有速度上的影响；针对第三、基于GCT最佳匹配的自适应色散补偿算法，周期性的信号在其时钟频率处有一个较强的频谱成分称之为GCT。GCT的幅度或者功率(幅度的平方)对信号中的残余色散极为敏感，可作为色散估计参数，基于GCT最佳匹配的色散估计方法在高速率的系统中对采样频偏敏感，同时现有的GCT最佳匹配色散估计方法不论采用色散扫描方式还是直接计算，复杂度都较高。
在接收端，长距离传输累积色度色散的宽脉冲响应可能分散在数百甚至上千个码元，且跟信号采样频率的平方成正比，为了实现效率的考虑，实现频域补偿时的快速傅里叶变换(Fast Fourier Transformation，简称FFT)长度很大，因为FFT长度较大，频谱较密，从而使得系统在频域处理时对采样频谱敏感。同时，因为FFT长度较大，在进行计算时复杂度很大，所需资源巨大。基于此，本申请发明人在研究中可选地发现，通过在频域上进行数据处理和滤波，展宽频点的频谱宽度，可以解决色散估计方法中对采样频谱敏感的问题，从而提供一种简单快速、高精度的色散估计方法，以及用于实现色散估计方法的色散估计装置、光通信数据相干接收机和存储介质。
请参阅图1，本申请一实施例提供了一种光通信数据相干接收机，包括依次连接的IQ不平衡补偿模块11、色散估计与补偿模块12、偏振解复用模块13、载波恢复模块14和判决译码模块15。
其中，IQ不平衡补偿模块11用于对IQ不平衡进行补偿。采用基带采样的数字通信接收器的模拟领域中，将射频或中频信号降频转换成基带后，将该基带信号通过模数转换器(Analog-to-Digital Converter，简称ADC)转换成数字信号。这种情况下，降频转换成基带时会被分离成同相信号与正交信号，此为藉由一个局部振荡器(Local Oscillator)使用增益相同且相位相差90度的两种正弦波(可选地讲为余弦、正弦波)达成。但是，由于这些过程是在模拟领域中执行，因此会产生误差。特别是，使用于降频转换的余弦波、正弦波相互之间会产生增益和相位误差，这时这些误差有可能会对接收机的性能产生严重影响，这叫IQ不平衡。
色散估计与补偿模块12用于使用接收机系统中色散补偿的数据流进行色散值估计，并将色散估计值用于系统的色散补偿。
偏振解复用模块13用于实现偏振解复用技术。偏振复用是指针对两束相同或不同波长的光可以同时在一根光纤中相互独立地传输，从而使得光纤的信息传输能力提高一倍且无需增加额外的带宽资源。而复用技术是指在信号发送端将多路信号按照不同的方法或区分方式进行组合，经历同一个信道传输后，在接收端将原本复用的信号分离出来，达到有效利用的目的。解复用方式主要包括直接检测和相干检测。
载波恢复模块14用于载波的恢复，从已调信号中恢复原载波。载波恢复的方法主要包括两种，一种是在发送端发送数字信号序列的同时也发送载波或与它有关的导频信号，在接收端可用窄带滤波器或锁相环直接提取载波；一种是接收信号为抑制载波的已调信号，通过对数字信号进行非线性变换或采用特殊的锁相环来获得相干载波。
判决译码模块15用于对纠错码实现最佳或者接近最佳的译码。判决译码主要包括软判决译码和硬判决译码，其中，软判决译码是指利用数字数据对纠错码实现译码，硬判决译码则是指译码器利用码的代数结构对纠错码进行译码。
请参阅图2，本申请一实施例提供了一种色散值估计方法，可应用于如图1所示的接收机中，具体的，本申请实施例所提供的色散值估计方法可应用于接 收机系统中的色散估计与补偿模块中的色散估计。该色散值估计方法包括如下步骤：
步骤101，获取频域数据，根据频域数据包含的正交的两个偏振态数据得到待处理数据；
这里，频域数据可以是采用色散补偿中已有的频域数据，也可以是将时域数据进行频域转换后得到的频域数据。正交的两个偏振态数据相互独立。接收机获取频域数据，根据频域数据包含的正交的两个偏振态数据得到待处理数据是指，接收机获取频域数据，对频域数据进行预处理，通过预处理将相互独立的两个偏振态数据分开，分别作为待处理数据；或者，通过预处理将相互独立的两个偏振态数据进行线性组合，形成多路分别包含两个偏振态分量的线性组合后的待处理数据。其中，将时域数据进行频域转换得到频域数据可以采用快速傅里叶变换(Fast Fourier Transition，简称FFT)来实现，两个偏振态数据可以分别用X[K]和Y[K]表示。
步骤103，对待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；
对待处理数据进行滤波的目的是指将频点对应的频谱范围展宽，以增加抵抗采样频偏的能力。将相邻频点之间的频谱宽度展宽的设置倍数可以通过配置滤波器系数实现，该设置倍数可以预先计算或仿真得到，也可以是根据经验值进行预先设置，如设置倍数可以为2倍或者3倍。需要说明的是，根据步骤101得到待处理数据通常包括两组或者两组以上，对待处理数据进行后续处理如滤波时，相应是指针对每一组待处理数据分别独立进行。
步骤105，对滤波后的待处理数据进行抽取得到目标数据组；
这里，对滤波后的待处理数据进行抽取是指将滤波后输出的数据划分为多组，并从多组中抽取其中一组。可以根据数据滤波中滤波器的系数的设置，对滤波后输出的待处理数据进行抽取。需要说明的是，根据步骤101得到待处理数据通常包括两组或者两组以上，对待处理数据进行后续处理，如滤波时是指针对每一组待处理数据分别独立进行，相应对滤波后的待处理数据进行抽取也是指针对每一组待处理数据分别独立进行。
步骤107，根据目标数据组计算分别与设置间隔值对应的目标函数，根据目标函数对应的相角值计算色散估计值。
设置间隔值可以包括多个，目标函数包括分别与设置间隔值对应的多个目标函数，接收机根据目标数据组计算分别与设置间隔值对应的目标函数，根据目标函数对应的相角值计算色散估计值，包括：接收机分别计算目标函数的相角值，并将多个设置间隔值分别对应的目标函数的相角值进行合并处理，得到最终的目标函数对应的相角值来计算色散估计值。
上述实施例中，该色散估计方法通过获取频域数据，根据频域数据包含的正交的两个偏振态数据得到待处理数据，对待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数，对滤波后的待处理数据进行抽取得到目标数据组，根据目标数据组计算分别与设置间隔值对应的目标函数，根据目标函数对应的相角值计算色散估计值，如此，通过在频域上对数据进行处理和滤波，展宽频点的频谱宽度，可以避免系统在频域处理时对采样频谱敏感的问题，实现精确色散估计，且通过根据目标函数对应的相角值计算色散估计值，简化了色散估计算法的计算量，以较简单的方法实现了色散值的准确估计，节省系统资源，降低系统功耗。
在一些实施例中，根据频域数据包含的正交的两个偏振态数据得到待处理数据，包括：
将频域数据中X偏振态数据和Y偏振态数据分别按照索引进行奇、偶分开，将偶数索引的频域数据取反且奇数索引的频域数据不变，得到预处理后的X偏振态数据和Y偏振态数据；
将预处理后的X偏振态数据和Y偏振态数据分别作为待处理数据；或者，将预处理后的X偏振态数据和Y偏振态数据进行线性组合运算，得到多组待处理数据。
这里，对频域数据进行预处理得到待处理数据包括：将获取的频域数据中X偏振态数据和Y偏振态数据分别按照索引进行奇、偶分开，将所有偶数索引的频域数据乘以-1且奇数索引的频域数据不变，得到预处理后的X偏振态数据和Y偏振态数据。以X偏振态数据为例，将X偏振态数据中偶数索引的频域数据乘以-1、且奇数索引的频域数据不变，得到预处理后的X偏振态数据，可以用如下公式表示：X′[K]＝X[K]*(-1) K，其中，K＝0，1，…N-1，N为频域数据长度。两个偏振态相互独立，将预处理后的X偏振态数据X[K]和Y 偏振态数据Y[K]分别作为两组待处理数据；或者，将预处理后的X偏振态数据X[K]和Y偏振态数据Y[K]进行线性组合运算，得到多个不同方向上的多组待处理数据。
在一些实施例中，将预处理后的X偏振态数据和Y偏振态数据进行线性组合运算，得到多组待处理数据，包括：
将X偏振态数据或Y偏振态数据作为第一待处理数据；
将X偏振态数据与Y偏振态数据取反后相加得到第二待处理数据；
将X偏振态数据和Y偏振态数据取反后分别作为复数的实部和虚部得到第三待处理数据。
这里，线性组合运算主要包括将相应偏振态数据进行取反、线性相加和复数相加的方式，从而得到三个不同方向上包含X偏振态数据和/或Y偏振态数据的三组待处理数据。具体的，第一待处理数据X 1[K]为X偏振态数据，或Y偏振态数据取反得到，以X偏振态数据为例，可以采用如下公获得：
X 1[K]＝X′[K]；
第二待处理数据X 2[K]可以将X偏振态数据与Y偏振态数据取反后线性相加得到，可以采用如下公式获得：
Y′[K]＝Y[K]*(-1) K；
X 2[K]＝X′[K]+Y′[K]；
第三待处理数据X 3[K]可以将X偏振态数据和Y偏振态数据取反后复数相加得到，可以采用如下公式获得：
X 3[K]＝X′[K]+jY′[K]；其中，j为虚数单位。
需要说明的是，上述将X偏振态数据和Y偏振态数据进行线性组合运算，得到多组待处理数据的计算过程中，X偏振态数据和Y偏振态数据可以互换，比如第一待处理数据可以为Y偏振态数据Y[K]，第三待处理数据可以将Y偏振态数据取反后作为复数的实部，将X偏振态数据作为复数的虚部。
上述实施例中，通过将X偏振态数据和Y偏振态数据进行线性组合运算，得到三个不同方向上的待处理数据分别执行色散值估算，可以避免偏振色散对色度色散估计的影响。
在一些实施例中，对待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数，包括：
根据频域数据长度和将相邻频点之间的频谱宽度预展宽的设置倍数确定窗函数；
根据频域数据对应的时域数据加窗函数处理所对应的频域卷积确定频域滤波系数，根据频域滤波系数对待处理数据进行滤波。
这里，频域滤波器系数根据频域数据对应的时域数据加指定长度的窗函数处理所对应的频域卷积进行确定。以频域数据长度为N、相邻频点之间的频谱宽度预展宽的设置倍数为M为例，根据频域数据长度和将相邻频点之间的频谱宽度预展宽的设置倍数确定窗函数包括，根据频域数据长度和将相邻频点之间的频谱宽度预展宽的设置倍数的比值N/M确定窗函数的长度，窗函数的类型可以根据具体情况进行选择，如可以为矩形窗、汉明窗等。
上述实施例中，将频域相邻频点之间的频域宽度展宽为原来的M倍，则滤波器系数为频域数据对应的时域数据加N/M长度的窗函数后对应的频域滤波器系数，也即，频域滤波后的效果等同于在时域上数据中间N/M的数据上加窗函数，从而实现将频点对应的频谱范围展宽至原来的M倍，以增加抵抗采样频偏的能力，在后续色散估计中，避免由于采样频偏的存在，对数据进行共轭相乘时频点偏移而导致目标函数误差增加或完全错误，通过增加两个频点之间的频谱宽度，可以解决频点偏移的问题。
在一些实施例中，请参阅图3，步骤107中，根据目标数据组计算分别与设置间隔值对应的目标函数，包括：
步骤1071，对目标数据组进行自相关；
步骤1072，将自相关后的目标数据组按照设置间隔值进行共轭相乘并累加，得到与设置间隔值对应的目标函数。
自相关是指信号在1个时刻的瞬时值与另1个时刻的瞬时值之间的依赖关系。设置间隔值的大小可以根据实际应用情况进行确定，通常，设置间隔值越小，则估计值范围越大，精度相应较低；反之，设置间隔值越大，则估计值范 围越小，精度相应较高。该设置间隔值的数量可以是多个，以得到分别与多个设置间隔值对应的目标函数。可选的，以设置间隔值为△ 1表示，以步骤101中对频域数据进行预处理得到的待处理数据中的X偏振态数据为例，将自相关后的目标数据组按照设置间隔值进行共轭相乘并累加，得到与设置间隔值对应的目标函数的公式(六)如下：
Figure PCTCN2019122107-appb-000001
其中，Rx[n]表示自相关后的数据，
Figure PCTCN2019122107-appb-000002
设置间隔值包括多组时，如△ 2,△ 3,...，可以得到对应的目标函数F 2,F 3,...，其中计算公式同上述公式(六)，在此不再赘述。
上述实施例中，通过对自相关的输出按照设置间隔值进行共轭相乘并累加得到与设置间隔值对应目标函数，从而可以根据实际应用情况确定设置间隔值的大小和数量，以调节估计值范围的精度。
在一些实施例中，步骤1071，对目标数据组进行自相关，包括：将目标数据组中频域上间隔波特率的数据进行共轭相乘。
目标数据组是指针对待处理数据进行抽取后得到的数据。对目标数据组进行自相关，即指将频域上间隔波特率的数据进行共轭相乘。仍以步骤101中对频域数据进行预处理得到的待处理数据分别为X偏振态数据X[K]和Y偏振态数据Y[K]为例，对待处理数据X[K]、Y[K]进行滤波、抽取得到的目标数据组分别以DX[L]、DY[L]表示，以X偏振态数据X′[K]为例，对目标数据组DX[L]进行自相关可以通过如下公式(七)实现：
Figure PCTCN2019122107-appb-000003
其中，
Figure PCTCN2019122107-appb-000004
conj(·)表示对数据取共轭。
在一些实施例中，在步骤107中，设置间隔值包括多个不同的设置间隔值，根据目标函数对应的相角值计算色散估计值，包括：
步骤1073，根据分别与不同的设置间隔值对应的目标函数的相角值之间的 关系是否符合设置条件进行合并，得到目标函数的目的相角值；
步骤1074，根据目的相角值、目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率计算色散估计值。
设置间隔值为多个且大小不同的的设置间隔值，设置间隔值越小，则估计值范围越大，精度相应较低；反之，设置间隔值越大，则估计值范围越小，精度相应较高，通过多个不同的设置间隔值，得到分别与多个设置间隔值对应的目标函数，并根据目标函数分别对应的相角值进行相位计算，根据相位计算输出的目的相角值乘以对应的系数得到色散估计值。其中，对应的系数由目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率确定，根据目的相角值、目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率计算色散估计值的计算公式如下公式八所示：
Figure PCTCN2019122107-appb-000005
其中，c为光速，单位为米/秒(m/s)；λ表示频域数据对应的光信号的中心波长，单位为纳米(nm)；nfft为离散傅里叶变换长度；f s表示采样频率，单位为GHz，f d表示符号速率，单位为Gbaud；最终输出的色散估计值CD的单位为纳秒/纳米(ns/nm)。
在一些实施例中，步骤1073，根据分别与不同的设置间隔值对应的目标函数的相角值之间的关系是否符合设置条件进行合并，得到目标函数的目的相角值，包括：
将相邻的两个设置间隔值对应的目标函数的相角值作为初始的第一相角值和第二相角值；
根据两个设置间隔值的比值和第一相角值的乘积得到对应的整数部分和小数部分；
当第二相角值与小数部分的差值大于或等于第一预设值时，将第二相角值与整数部分合并后减1作为更新后的第二相角值；
当第二相角值与小数部分的差值小于第二预设值时，将第二相角值与整数部分合并后加1作为更新后的第二相角值；
当第二相角值与小数部分的差值小于第一预设值且大于或等于第二预设值时，将第二相角值与整数部分合并作为更新后的第二相角值；
将更新后的第二相角值、及依序选取下一设置间隔值对应的目标函数的相角值作为初始的第一相角值和第二相角值，返回执行上述步骤，直至设置间隔值对应的目标函数的相角值完成合并，得到目标函数的目的相角值。
这里，假设设置间隔值的数量为w(w>1)个，且△ 1<△ 2<...<△ w。其中，设置间隔值△ 1对应的目标函数F1的相角值为
Figure PCTCN2019122107-appb-000006
设置间隔值△ 2对应的目标函数F2的相角值为
Figure PCTCN2019122107-appb-000007
以此类推，设置间隔值△ w对应的目标函数F3的相角值为
Figure PCTCN2019122107-appb-000008
以第一预设值为0.5，第二预设值为-0.5为例，根据分别与不同的设置间隔值对应的目标函数的相角值之间的关系是否符合设置条件，对
Figure PCTCN2019122107-appb-000009
进行合并如下公式(九)至(十一)所示：
Figure PCTCN2019122107-appb-000010
Figure PCTCN2019122107-appb-000011
Figure PCTCN2019122107-appb-000012
如果
Figure PCTCN2019122107-appb-000013
如果
Figure PCTCN2019122107-appb-000014
其他情况则
Figure PCTCN2019122107-appb-000015
Figure PCTCN2019122107-appb-000016
其中，angle(·)表示取数据的相角，floor(·)表示向下取整，从而上述公式十中，itr表示根据两个设置间隔值的比值和第一相角值的乘积得到的整数部分、u表示小数部分；n＝2，3，…w。
其中，n取值为2时，则
Figure PCTCN2019122107-appb-000017
和
Figure PCTCN2019122107-appb-000018
分别为
Figure PCTCN2019122107-appb-000019
和
Figure PCTCN2019122107-appb-000020
也即，将
Figure PCTCN2019122107-appb-000021
和
Figure PCTCN2019122107-appb-000022
作为初始的第一相角值和第二相角值根据公式(九)至(十一)执行合并，得到更新后的
Figure PCTCN2019122107-appb-000023
再将更新后的
Figure PCTCN2019122107-appb-000024
以及依序选取的
Figure PCTCN2019122107-appb-000025
(也即n取值为3时)为初始的第一相角值和第二相角值根据公式(九)至(十一)执行合并，得到更新后的
Figure PCTCN2019122107-appb-000026
依次类推，设置间隔值对应的目标函数的相角值完成合并，得到目标函数的目的相角值
Figure PCTCN2019122107-appb-000027
在一些实施例中，步骤1074，根据目的相角值、目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率计算色散估计值之前，包括：
对最近的多次计算的目标函数的目的相角值进行比较，确定任意两个之差的绝对值小于门限值时，则将当前得到的目的相角值作为最终的目的相角值；
确定任意两个之差的绝对值大于门限值时，则将前一次得到的目的相角值作为最终的目的相角值。
这里，最近的多次计算的目标函数的目的相角值可以是指截止至当前时间的最近的多次计算的NL个目标函数的目的相角值，其中，NL的值可以根据实际需求进行设置。当这NL个目的相角值之间任意两个之差的绝对值小于门限值时，则选取当前计算得到的目的相角值作为最终的目的相角值，从而相应是以根据当前计算得到的目的相角值计算色散估计值；当这NL个目的相角值之间任意两个之差的绝对值大于门限值时，则将前一次得到的目的相角值作为最终的目的相角值，从而相应是以该前一次得到的目的相角值计算色散估计值。
在一些实施例中，根据目标函数对应的相角值计算色散估计值之前，包括：
对目标函数进行平滑滤波。
这里，对目标函数进行平滑滤波的滤波器系数采用低通滤波系数。以目标函数为F 1,F 2,F 3,...F w例，进行平滑滤波后得到的滤波输出为F 1',F 2',F 3',...F w'。
在一些实施例中，根据目标函数对应的相角值计算色散估计值之前，包括：
构造多个设置长度的寄存器，将最近的多次计算的目标函数的相角值对应 存放至寄存器内，根据每一寄存器的累加结果得到目标函数对应的相角值。
这里，寄存器的设置长度与存放至寄存器内的最近的多次计算的目标函数的相角值的计算次数相同。寄存器的数量与目标函数的数量对应。以目标函数为F 1,F 2,F 3,...F w、将最近的NL次计算的目标函数的相角值对应存放至寄存器为例，构造W个长度为NL的寄存器，用buffer1，buffer2，...，buffer W表示，分别存放最近NL次计算所得目标函数F 1,F 2,F 3,...F w。buffer1-W中初始值都是NL个0，buffer1-W中的累加结果对应的目标函数为F 1_sum,F 2_sum,F 3_sum,...F w_sum，再根据buffer1-W中的累加结果得到的目标函数F 1_sum,F 2_sum,F 3_sum,...F w_sum计算目标函数的目的相角值。
在另一可选的实施例中，对最近的多次计算的目标函数的目的相角值进行比较，可以是根据buffer1-W中的累加结果对应的目标函数F 1_sum,F 2_sum,F 3_sum,...F w_sum的相角值计算得到的最近的NL个目的相角值进行比较，当这NL个目的相角值之间任意两个之差的绝对值小于门限值时，则选取当前计算得到的目的相角值
Figure PCTCN2019122107-appb-000028
作为最终的目的相角值；当这NL个目的相角值之间任意两个之差的绝对值大于门限值时，则将前一次得到的目的相角值
Figure PCTCN2019122107-appb-000029
作为最终的目的相角值。
在一些实施例中，根据目标函数对应的相角值计算色散估计值之后，包括：
对最近的多次计算的目标函数的目的相角值进行比较，确定任意两个之差的绝对值小于门限值时，则根据当前得到的色散估计值更新色散估计值；
确定任意两个之差的绝对值大于门限值时，则保持前一次得到的色散估计值。
这里，最近的多次计算的目标函数的目的相角值可以是指截止至当前时间的最近的多次计算的NL个目标函数的目的相角值，其中，NL的值可以根据实际需求进行设置。当这NL个目的相角值之间任意两个之差的绝对值小于门限值时，才开始更新色散估计值；反之，则停止更新色散值估计值，保持上次的色散估计值。
在一些实施例中，步骤105，对滤波后的待处理数据进行抽取得到目标数据组，包括：
将滤波后的待处理数据划分为与设置倍数数量相同的数据组，从数据组抽取得到目标数据组。
这里，数据抽取与数据滤波中滤波器系数的设置相关。若将频域中两个频点之间的频谱宽度变为原来的M倍，则将滤波后的待处理数据划分为M组，以间隔M的数据作为一组。该待处理数据划分为M组后得到的数据组可以表示如下：
索引为L·M+1的作为第一组，索引为L·M+2的作为第二组，以此类推，索引为L·M+M的作为第M组。其中
Figure PCTCN2019122107-appb-000030
从数据组抽取得到目标数据组可以是，选取M组中的第P组作为目标数据组。其中P可以预先配置且设置为1至M中的任意数值。
为了能够对本申请实施例所提供的色散估计方法的实现流程能够进一步具体的了解，下面分别对色散估计方法的四个可选的实施例进行说明，请参阅图4，该色散估计方法包括如下步骤：
步骤S11，数据预处理，将X偏振态数据和Y偏振态数据分别作为待处理数据；
获取频域数据，通过数据预处理将获取的频域数据按照索引进行奇、偶分开，将所有偶数索引的频域数据乘以-1，奇数索引的频域数据不变。高速超高速光纤通信中数据分为X偏振态和Y偏振态这两个相交的偏振态，两个偏振态相互独立。以X偏振态数据为例，Y偏振态数据操作相同。将X偏振态数据进行预处理，是指通过如下方式进行处理：X′[K]＝X[K]·(-1) k，其中，K＝0，1，2…N-1。N为频域数据长度。
步骤S12，对数据预处理后输出的待处理数据进行滤波并输出；
其中，针对预处理后的X偏振态数据X[K]和Y偏振态数据Y[K]分开进行FIR滤波。滤波的目的是将频点对应的频谱范围展宽，进而增加抵抗采样频偏的能力。在色散估计中，后续需要对频域数据在频域上间隔波特率的数据进行共轭相乘操作。因为采样频偏的存在，可能会造成共轭相乘时频点偏移，这会照成 目标函数误差增大或者完全错误。通过增加两个频点之间的频谱宽度，可以避免频点偏移问题。滤波器系数可以根据系统具体情况配置，具体的，需要将频点展宽为原来的几倍可以事先计算或者仿真得到，这个值也可以根据经验一般为2倍或者3倍。若将频域两个频点之间的频谱宽度变为原来的M倍，则滤波器系数为时域加N/M长度窗后对应的频域滤波器系数，即频域滤波后的效果等同于在时域上数据中间N/M的数据上加窗函数，窗函数可以根据具体情况进行选择，例如矩形窗，汉明窗等。N为频域数据长度，M为大于等于2的整数。
步骤S13，对数据滤波输出的数据进行抽取；
根据数据滤波中滤波器系数的设置，对数据滤波输出进行数据抽取。X偏振态数据和Y偏振态数据分开进行数据抽取。
若将频域两个频点之间的频谱宽度变为原来的M倍，则将数据滤波输出的数据划分为M组，以间隔M的数据作为一组。具体的，索引为L·M+1的作为第一组，索引为L·M+2的作为第二组，以此类推，索引为L·M+M的作为第M组。
Figure PCTCN2019122107-appb-000031
选取M组当中的第P组作为输出，P可以设置为1至M中的任意数值。需要注意，一旦选择第P组作为输出，则后面所有数据抽取操作均要选择第P组作为输出。P作为参数只能在系统开始或者重启时进行配置，其他情况不能变化。
步骤S14，对抽取后的数据进行自相关操作；
将频域上间隔波特率的数据进行共轭相乘。假设对X偏振态数据和Y偏振态数据进行数据抽取的输出分别为DX[L]，DY[L]，
Figure PCTCN2019122107-appb-000032
使用X偏振态数据为例，使用Y偏振态数据效果相同，对抽取后的数据进行自相关操作的计算公式如下：
Figure PCTCN2019122107-appb-000033
其中
Figure PCTCN2019122107-appb-000034
conj(·)表示对数据取共轭。
步骤S15，对进行自相关操作后的数据按照设置间隔值进行共轭相乘并累加，得到目标函数；
按照事先设置的间隔值△ 1计算得到目标函数：
Figure PCTCN2019122107-appb-000035
设置的间隔值可以有多组，例如△ 2,△ 3,...，从而可以得到对应的目标函数F 2,F 3,...。
S16，根据目标函数对应的相角值进行相位计算，得到最终的目标函数相角值；
计算目标函数对应的相角值，将不同间隔值的相角值进行合并处理，假设有W(W>1)个设置间隔值，且△ 1<△ 2<...<△ w。
△ 1对应的相角值
Figure PCTCN2019122107-appb-000036
△ 2对应的相角值
Figure PCTCN2019122107-appb-000037
以此类推，△ w对应的相角值
Figure PCTCN2019122107-appb-000038
对
Figure PCTCN2019122107-appb-000039
进行如下合并操作：
Figure PCTCN2019122107-appb-000040
Figure PCTCN2019122107-appb-000041
Figure PCTCN2019122107-appb-000042
如果
Figure PCTCN2019122107-appb-000043
如果
Figure PCTCN2019122107-appb-000044
其他情况
Figure PCTCN2019122107-appb-000045
Figure PCTCN2019122107-appb-000046
其中，
Figure PCTCN2019122107-appb-000047
即为最终的目标函数相位值。
此处，n＝2,3,...,w，其中angle(·)表示取数据的相角，floor(·)表示向下取整。
S17，根据相位计算输出的目标函数相角值计算色散估计值并输出；
将相位计算输出的相角值乘以对应的系数得到色散估计值，并将色散估计值输出，其中，色散估计值的计算公式如下：
Figure PCTCN2019122107-appb-000048
其中，c为光速，单位为米/秒(m/s)；λ表示频域数据对应的光信号的中心 波长，单位为纳米(nm)；nfft为离散傅里叶变换长度，f s表示采样频率，单位为GHz，f d表示符号速率，单位为Gbaud；最终输出的色散值CD的单位为纳秒/纳米(ns/nm)。
请参阅图5，在另一实施例中，该色散估计方法包括如下步骤：
步骤S21，数据预处理，将X偏振态数据和Y偏振态数据进行线性组合得到三组待处理数据；
获取频域数据，通过数据预处理将获取的频域数据按照索引进行奇、偶分开，将所有偶数索引的频域数据乘以-1，奇数索引的频域数据不变。高速超高速光纤通信中数据分为X偏振态和Y偏振态这两个相交的偏振态，两个偏振态相互独立。以X偏振态数据为例，将X偏振态数据中偶数索引的频域数据乘以-1且奇数索引的频域数据不变，得到预处理后的X偏振态数据，其中预处理可以用如下公式表示：X′[K]＝X[K]*(-1) K，其中，K＝0，1，2…N-1，N为频域数据长度；Y偏振态数据的处理方式与X偏振态数据的预处理方式相同，在此不再赘述。
将预处理后的X偏振态数据和Y偏振态数据进行如下线性组合运算：
X 1[k]＝X'[k]
X 2[k]＝X'[k]+Y'[k]
X 3[k]＝X'[k]+j·Y'[k]
其中，Y′[K]＝Y[K]·(-1) k，其中，K＝0，1，2…N-1，N为频域数据长度；j为虚数单位，将线性组合运算得到的X 1[k],X 2[k],X 3[k]作为数据预处理输出的三组待处理数据。
步骤S22，对数据预处理后输出的待处理数据进行滤波并输出；与步骤S12的不同在于，将滤波输入的数据替换为X 1[k],X 2[k],X 3[k]，相应的，数据滤波的输出结果也相应变为三组。
步骤S23，对数据滤波输出的数据进行抽取；与步骤S13的不同在于，将抽取输入的数据替换为步骤S22的三组输出结果，抽取输出结果也变为三组；
步骤S24，对抽取后的数据进行自相关操作；与步骤S14的不同在于，将自相关操作输入的数据替换为步骤S23的三组抽取输出结果；
步骤S25，对进行自相关操作后的数据按照设置间隔值进行共轭相乘并累加，得到目标函数；与步骤S15的不同在于，根据每一组输入分别计算目标函数，得到三组目标函数；
三组目标函数可以表示为：F 1_1,F 2_1,F 3_1,...F w_1、F 1_2,F 2_2,F 3_2,...F w_2、F 1_3,F 2_3,F 3_3,...F w_3，针对每一组输入进行共轭相乘的计算方式与步骤S15相同，其中，根据三组目标函数相应累加的计算公式如下：
F′ 1＝F 1_1+F 2_1+F 3_1
F′ 2＝F 2_1+F 2_1+F 2_1
F′ 3＝F 3_1+F 3_1+F 3_1
将F′ 1,F′ 2,F′ 3,...F′ w作为目标函数输出。
步骤S26，根据目标函数对应的相角值进行相位计算，得到最终的目标函数相角值；与步骤S16的不同在于，将相位计算的输入替换为步骤S25输出的目标函数F′ 1,F′ 2,F′ 3,...F′ w。
步骤S27，根据相位计算输出的目标函数相角值计算色散估计值并输出；与步骤S17相同。
请参阅图6，在又一实施例中，该色散估计方法包括如下步骤：
步骤S31，数据预处理，将X偏振态数据和Y偏振态数据分别作为待处理数据；与步骤S11相同。
步骤S32，对数据预处理后输出的待处理数据进行滤波并输出；与步骤S12相同。
步骤S33，对数据滤波输出的数据进行抽取；与步骤S13相同。
步骤S34，对抽取后的数据进行自相关操作；与步骤S14相同。
步骤S35，对进行自相关操作后的数据按照设置间隔值进行共轭相乘并累加，得到目标函数后进行平滑滤波，输出滤波后的目标函数；与步骤S15的不同在于：对按照多组设置的间隔值，如△ 2,△ 3,...△ w计算得到的目标函数F 2,F 3,...F w，进行平滑滤波，滤波器系数采用低通滤波系数，滤波输出为 F′ 1,F′ 2,F′ 3,...F′ w；
步骤S36，根据滤波后的目标函数对应的相角值进行相位计算，得到最终的目标函数相角值；与步骤S16的不同在于：将相位计算的输入替换为步骤S35输出的目标函数F′ 1,F′ 2,F′ 3,...F′ w。
步骤S37，根据相位计算输出的目标函数相角值计算色散估计值，根据最近的多个相角值之间任意两个之差的绝对值与门限值的关系，对色散估计值进行更新；与步骤S17的不同在于：根据相位计算输出的目标函数相角值计算色散估计值后，对最近的多个相角值
Figure PCTCN2019122107-appb-000049
进行比较，如果多个相角值
Figure PCTCN2019122107-appb-000050
之间任意两个之差的绝对值小于门限值Th，才开始更新色散估计值，将当前计算得到的色散估计值输出；如果多个相角值
Figure PCTCN2019122107-appb-000051
之间任意两个之差的绝对值大于等于门限值Th，则停止更新色散估计值，保持上次的色散估计值，将上次的色散估计值输出。
请参阅图7，在再一实施例中，该色散估计方法包括如下步骤：
步骤S41，数据预处理，将X偏振态数据和Y偏振态数据分别作为待处理数据；与步骤S11相同。
步骤S42，对数据预处理后输出的待处理数据进行滤波并输出；与步骤S12相同。
步骤S43，对数据滤波输出的数据进行抽取；与步骤S13相同。
步骤S44，对抽取后的数据进行自相关操作；与步骤S14相同。
步骤S45，对进行自相关操作后的数据按照设置间隔值进行共轭相乘并累加，得到目标函数；与步骤S15相同。
步骤S46，构造多个长度为NL的寄存器存放最近NL次计算得到的目标函数，根据寄存器中数据分别累加得到累加后的目标函数，根据累加后的目标函数对应的相角值进行相位计算，得到最终的目标函数相角值；与步骤S16的不同在于：
构造W个长度为NL的buffer1，buffer2，...，buffer W，分别存放最近NL次计算所得的目标函数F 1,F 2,F 3,...F w。buffer1-W中初始值都是NL个0，根据寄存器中数据分别累加得到累加后的目标函数是指将buffer1-W中数据分别求 和得到F 1_sum,F 2_sum,F 3_sum,...F w_sum。
根据累加后的目标函数对应的相角值进行相位计算是指，根据buffer1-W中数据分别求和得到的目标函数F 1_sum,F 2_sum,F 3_sum,...F w_sum分别计算目标函数相角值
Figure PCTCN2019122107-appb-000052
其中计算过程与步骤S16相同。
步骤S461，根据寄存器中数据计算出的多个目标函数相角值之间任意两个之差的绝对值与门限值的关系，对最终的目标函数相角值进行更新；与步骤S16的不同在于：
在根据累加后的目标函数对应的相角值进行相位计算，得到最终的目标函数相角值之后，还包括：根据buffer1-W中数据计算出的多个目标函数相角值
Figure PCTCN2019122107-appb-000053
之间任意两个之差的绝对值小于门限值Th时，则将本次步骤S46计算得出的
Figure PCTCN2019122107-appb-000054
作为相位计算输出；其他情况，使用上次的
Figure PCTCN2019122107-appb-000055
作为相位计算输出。
其中
Figure PCTCN2019122107-appb-000056
为使用F 1_sum,F 2_sum,F 3_sum,...F w_sum计算得出的目标函数相位值。
步骤S47，根据相位计算输出的目标函数相角值计算色散估计值并输出；与步骤S17的不同在于：将相位计算输出的目标函数替换为根据步骤S461输出的目标函数相位值，其中色散估计值的计算方式与步骤S17相同。
相关技术中，在接收端，长距离传输累积的色散的宽脉冲响应可能分散在数百甚至上千个码元，且跟信号采样频率的平方成正比。为了实现效率的考虑，实现频域补偿时的FFT长度会很大，所以频谱较密，从而使得系统在频域处理时对采样频谱敏感，所以传统的色散估计方法在超高速光纤通信系统中不能正常工作。本申请上述实施例提供的色散估计方法，解决了接收端通过在频域上进行数据预处理和滤波，展宽频点的频谱宽度，从而解决了超高速长途光纤系统色散估计方法对采样频偏敏感的问题，实现了超高速光纤通信的精确色散估计。同时通过计算目标函数值对应的相角值，并将不同间隔时的相位值进行拼接合并处理，得到最终的目标函数相位值，进而计算出接收机系统的色散估计值。与相关技术提供的基于色散扫描操作或者FFT操作的色散估计方法相比，简化了色散估计算法的计算量，以较简单的方法实现了超高速长途光纤通信系统的色散值的准确估计，节省了系统资源，降低了系统功耗。
如表1所示，是采用本申请实施例所提供的色散估计方法计算一次色散估计值所需系统资源与相关技术中基于FFT操作的色散估计方法的对比表，以数据位宽10位为例，采用本申请实施例所提供的色散估计方法根据目标函数值计算一次色散估计值的计算量相比基于传统FFT法计算一次色散估计值相比节省了约78.6％的资源。
表1
Figure PCTCN2019122107-appb-000057
本申请实施例另一方面，请参阅图8，还提供一种色散估计装置，包括频域数据获取模块121，用于获取频域数据，根据频域数据包含的正交的两个偏振态数据得到待处理数据；滤波模块123，用于对待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；抽取模块125，用于对滤波后的待处理数据进行抽取得到目标数据组；色散估计模块127，用于根据目标数据组计算分别与设置间隔值对应的目标函数，根据目标函数对应的相角值计算色散估计值。
在一些实施例中，频域数据获取模块121，包括偏振态数据单元、第一预处理单元或第二预处理单元，偏振态数据单元，用于将频域数据中X偏振态数据和Y偏振态数据分别按照索引进行奇、偶分开，将偶数索引的频域数据取反且奇数索引的频域数据不变，得到预处理后的X偏振态数据和Y偏振态数据；第一预处理单元，用于将预处理后的X偏振态数据和Y偏振态数据分别作为待处理数据；第二预处理单元，用于将预处理后的X偏振态数据和Y偏振态数据进行线性组合运算，得到多组待处理数据。
在一些实施例中，第二预处理单元，具体用于将X偏振态数据或Y偏振态数据作为第一待处理数据；将X偏振态数据与Y偏振态数据取反后相加得到第二待处理数据；将X偏振态数据和Y偏振态数据取反后分别作为复数的实部和虚部得到第三待处理数据。
在一些实施例中，滤波模块123，具体用于根据频域数据长度和将相邻频点之间的频谱宽度预展宽的设置倍数确定窗函数；根据频域数据对应的时域数据加窗函数处理所对应的频域卷积确定频域滤波系数，根据频域滤波系数对待处理数据进行滤波。
在一些实施例中，色散估计模块127包括自相关单元1271和目标函数获取单元1272，自相关单元1271，用于对目标数据组进行自相关；目标函数获取单元1272，用于将自相关后的目标数据组按照设置间隔值进行共轭相乘并累加，得到与设置间隔值对应的目标函数。
在一些实施例中，自相关单元1271，具体用于将目标数据组中频域上间隔波特率的数据进行共轭相乘。
在一些实施例中，设置间隔值包括多个不同的设置间隔值，色散估计模块127还包括相位计算单元1273和色散估计单元1274，相位计算单元1273，具体用于根据分别与不同的设置间隔值对应的目标函数的相角值之间的关系是否符合设置条件进行合并，得到目标函数的目的相角值；色散估计单元1274，用于根据目的相角值、目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率计算色散估计值。
在一些实施例中，相位计算单元1273，具体用于将相邻的两个设置间隔值对应的目标函数的相角值作为初始的第一相角值和第二相角值；根据两个设置间隔值的比值和第一相角值的乘积得到对应的整数部分和小数部分；当第二相角值与小数部分的差值大于或等于第一预设值时，将第二相角值与整数部分合并后减1作为更新后的第二相角值；当第二相角值与小数部分的差值小于第二预设值时，将第二相角值与整数部分合并后加1作为更新后的第二相角值；当第二相角值与小数部分的差值小于第一预设值且大于或等于第二预设值时，将第二相角值与整数部分合并作为更新后的第二相角值；将更新后的第二相角值、及依序选取下一设置间隔值对应的目标函数的相角值作为初始的第一相角值和第二相角值，返回执行上述步骤，直至设置间隔值对应的目标函数的相角值完成合并，得到目标函数的目的相角值。
在一些实施例中，相位计算单元1273，具体用于对最近的多次计算的目标函数的目的相角值进行比较，确定任意两个之差的绝对值小于门限值时，则将当前得到的目的相角值作为最终的目的相角值；确定任意两个之差的绝对值大 于门限值时，则将前一次得到的目的相角值作为最终的目的相角值。
在一些实施例中，目标函数获取单元1272，还用于在根据目标函数对应的相角值计算色散估计值之前，对目标函数进行平滑滤波。
在一些实施例中，相位计算单元1273，还用于在根据目标函数对应的相角值计算色散估计值之前，构造多个设置长度的寄存器，将最近的多次计算的目标函数的相角值对应存放至寄存器内，根据每一寄存器的累加结果得到目标函数对应的相角值。
在一些实施例中，色散估计单元1274，还用于在根据目标函数对应的相角值计算色散估计值之后，对最近的多次计算的目标函数的目的相角值进行比较，确定任意两个之差的绝对值小于门限值时，则根据当前得到的色散估计值更新色散估计值；确定任意两个之差的绝对值大于门限值时，则保持前一次得到的色散估计值。
在一些实施例中，抽取模块125，具体用于将滤波后的待处理数据划分为与设置倍数数量相同的数据组，从数据组抽取得到目标数据组。
上述实施例提供的色散估计装置在进行计算色散估计值时，仅以上述各程序模块的划分进行举例说明，实际应用中，可以根据需要而将上述步骤分配由不同的程序模块完成，即可以将装置的内部结构划分成不同的程序模块，以完成以上描述的全部或者部分处理。另外，上述实施例提供的色散估计装置与色散估计方法实施例属于同一构思，其具体实现过程详见方法实施例，这里不再赘述。
本申请实施例还提供了一种计算机存储介质，例如包括存储有计算机程序的存储器，该计算机程序可以由处理器执行，以完成本申请任一实施例所提供的色散估计方法的步骤。该计算机存储介质可以是FRAM、ROM、PROM、EPROM、EEPROM、Flash Memory、磁表面存储器、光盘、或CD-ROM等存储器；也可以是包括上述存储器之一或任意组合的各种设备。
以上，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。本申请的保护范围应以权利要求的保护范围以准。
工业实用性
通过本申请实施例所提供的色散估计方法、装置、光通信数据相干接收机及存储介质，利用获取频域数据，根据上述频域数据包含的正交的两个偏振态数据得到待处理数据，对上述待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数，对滤波后的上述待处理数据进行抽取得到目标数据组，根据上述目标数据组计算分别与设置间隔值对应的目标函数，根据上述目标函数对应的相角值计算色散估计值，如此，通过在频域上对数据进行处理和滤波，展宽频点的频谱宽度，可以避免系统在频域处理时对采样频谱敏感的问题，实现精确色散估计，且通过根据目标函数对应的相角值计算色散估计值，简化了色散估计算法的计算量，以较简单的方法实现了色散值的准确估计，节省系统资源，降低系统功耗。

Claims (16)

  1. 一种色散估计方法，包括：
    获取频域数据，根据所述频域数据包含的正交的两个偏振态数据得到待处理数据；
    对所述待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；
    对滤波后的所述待处理数据进行抽取得到目标数据组；
    根据所述目标数据组计算分别与设置间隔值对应的目标函数，根据所述目标函数对应的相角值计算色散估计值。
  2. 如权利要求1所述的色散估计方法，其中：所述根据所述频域数据包含的正交的两个偏振态数据得到待处理数据，包括：
    将所述频域数据中X偏振态数据和Y偏振态数据分别按照索引进行奇、偶分开，将所述偶数索引的频域数据取反且所述奇数索引的频域数据不变，得到预处理后的所述X偏振态数据和所述Y偏振态数据；
    将预处理后的所述X偏振态数据和所述Y偏振态数据分别作为待处理数据；或者，将预处理后的所述X偏振态数据和所述Y偏振态数据进行线性组合运算，得到多组待处理数据。
  3. 如权利要求2所述的色散估计方法，其中，所述将预处理后的所述X偏振态数据和所述Y偏振态数据进行线性组合运算，得到多组待处理数据，包括：
    将所述X偏振态数据或Y偏振态数据作为第一待处理数据；
    将所述X偏振态数据与所述Y偏振态数据取反后相加得到第二待处理数据；
    将所述X偏振态数据和所述Y偏振态数据取反后分别作为复数的实部和虚部得到第三待处理数据。
  4. 如权利要求1所述的色散估计方法，其中，所述对所述待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数，包括：
    根据所述频域数据长度和将相邻频点之间的频谱宽度预展宽的设置倍数确定窗函数；
    根据所述频域数据对应的时域数据加所述窗函数处理所对应的频域卷积确 定频域滤波系数，根据所述频域滤波系数对所述待处理数据进行滤波。
  5. 如权利要求1所述的色散估计方法，其中，所述根据所述目标数据组计算分别与设置间隔值对应的目标函数，包括：
    对所述目标数据组进行自相关；
    将自相关后的所述目标数据组按照设置间隔值进行共轭相乘并累加，得到与所述设置间隔值对应的目标函数。
  6. 如权利要求5所述的色散估计方法，其中，所述对所述目标数据组进行自相关，包括：
    将所述目标数据组中频域上间隔波特率的数据进行共轭相乘。
  7. 如权利要求1所述的色散估计方法，其中，所述设置间隔值包括多个不同的设置间隔值，所述根据所述目标函数对应的相角值计算色散估计值，包括：
    根据分别与不同的设置间隔值对应的目标函数的相角值之间的关系是否符合设置条件进行合并，得到所述目标函数的目的相角值；
    根据所述目的相角值、所述目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率计算色散估计值。
  8. 如权利要求7所述的色散估计方法，其中，所述根据分别与不同的设置间隔值对应的目标函数的相角值之间的关系是否符合设置条件进行合并，得到所述目标函数的目的相角值，包括：
    将相邻的两个设置间隔值对应的目标函数的相角值作为初始的第一相角值和第二相角值；
    根据所述两个设置间隔值的比值和所述第一相角值的乘积得到对应的整数部分和小数部分；
    当所述第二相角值与所述小数部分的差值大于或等于第一预设值时，将所述第二相角值与所述整数部分合并后减1作为更新后的第二相角值；
    当所述第二相角值与所述小数部分的差值小于第二预设值时，将所述第二相角值与所述整数部分合并后加1作为更新后的第二相角值；
    当所述第二相角值与所述小数部分的差值小于所述第一预设值且大于或等于所述第二预设值时，将所述第二相角值与所述整数部分合并作为更新后的第二相角值；
    将所述更新后的第二相角值、及依序选取下一设置间隔值对应的目标函数的相角值作为初始的第一相角值和第二相角值，返回执行上述步骤，直至所述设置间隔值对应的目标函数的相角值完成合并，得到所述目标函数的目的相角值。
  9. 如权利要求7所述的色散估计方法，其中，所述根据所述目的相角值、所述目的相角值对应的设置间隔值、光速、光信号的中心波长、采样频率、离散傅里叶变换长度和符号速率计算色散估计值之前，包括：
    对最近的多次计算的所述目标函数的目的相角值进行比较，确定任意两个之差的绝对值小于门限值时，则将当前得到的所述目的相角值作为最终的目的相角值；
    确定任意两个之差的绝对值大于门限值时，则将前一次得到的所述目的相角值作为最终的目的相角值。
  10. 如权利要求1所述的色散估计方法，其中，所述根据所述目标函数对应的相角值计算色散估计值之前，包括：
    对所述目标函数进行平滑滤波。
  11. 如权利要求1所述的色散估计方法，其中，所述根据所述目标函数对应的相角值计算色散估计值之前，包括：
    构造多个设置长度的寄存器，将最近的多次计算的所述目标函数的相角值对应存放至所述寄存器内，根据每一寄存器的累加结果得到所述目标函数对应的相角值。
  12. 如权利要求1所述的色散估计方法，其中，所述根据所述目标函数对应的相角值计算色散估计值之后，包括：
    对最近的多次计算的所述目标函数的目的相角值进行比较，确定任意两个之差的绝对值小于门限值时，则根据当前得到的所述色散估计值更新色散估计值；
    确定任意两个之差的绝对值大于门限值时，则保持前一次得到的所述色散估计值。
  13. 如权利要求1所述的色散估计方法，其中，所述对滤波后的所述待处理数据进行抽取得到目标数据组，包括：
    将滤波后的所述待处理数据划分为与所述设置倍数数量相同的数据组，从所述数据组抽取得到目标数据组。
  14. 一种色散估计装置，包括：
    频域数据获取模块，设置为获取频域数据，根据所述频域数据包含的正交的两个偏振态数据得到待处理数据；
    滤波模块，设置为对所述待处理数据进行滤波，将相邻频点之间的频谱宽度展宽设置倍数；
    抽取模块，设置为对滤波后的所述待处理数据进行抽取得到目标数据组；
    色散估计模块，设置为根据所述目标数据组计算分别与设置间隔值对应的目标函数，根据所述目标函数对应的相角值计算色散估计值。
  15. 一种光通信数据相干接收机，所述接收机包括依次连接的IQ不平衡补偿模块、色散估计与补偿模块、偏振解复用模块、载波恢复模块和判决译码模块，其中，所述色散估计与补偿模块用于实现权利要求1至13中任一项所述的色散估计方法。
  16. 一种存储介质，所述存储介质中存储有可执行指令，所述可执行指令被处理器执行时实现权利要求1至13中任一项所述的色散估计方法。
PCT/CN2019/122107 2018-11-30 2019-11-29 色散估计方法、装置、接收机及存储介质 WO2020108632A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
JP2021531064A JP7482128B2 (ja) 2018-11-30 2019-11-29 分散推定方法、装置、受信機及び記憶媒体
EP19888798.6A EP3890211A4 (en) 2018-11-30 2019-11-29 DISPERSION ESTIMATION METHOD, DEVICE, RECEIVER AND STORAGE MEDIA
KR1020217019935A KR20210096191A (ko) 2018-11-30 2019-11-29 색분산 추정 방법, 장치, 수신기 및 기록매체

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811449784.0 2018-11-30
CN201811449784.0A CN111262634B (zh) 2018-11-30 2018-11-30 色散估计方法、装置、接收机及存储介质

Publications (1)

Publication Number Publication Date
WO2020108632A1 true WO2020108632A1 (zh) 2020-06-04

Family

ID=70852651

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/122107 WO2020108632A1 (zh) 2018-11-30 2019-11-29 色散估计方法、装置、接收机及存储介质

Country Status (5)

Country Link
EP (1) EP3890211A4 (zh)
JP (1) JP7482128B2 (zh)
KR (1) KR20210096191A (zh)
CN (1) CN111262634B (zh)
WO (1) WO2020108632A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114205204B (zh) * 2020-09-02 2024-09-24 中兴通讯股份有限公司 频域广义线性均衡方法、装置、系统及非易失性存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6204924B1 (en) * 1999-02-23 2001-03-20 Exfo Electro-Optical Engineering Inc. Method and apparatus for measuring polarization mode dispersion of optical devices
EP1318622A2 (en) * 2001-12-05 2003-06-11 Nortel Networks Limited Polarisation mode dispersion measurement and compensation
CN104579476A (zh) * 2013-10-22 2015-04-29 中兴通讯股份有限公司 光相干通信中色散估计方法及装置
CN104780131A (zh) * 2014-01-15 2015-07-15 深圳市中兴微电子技术有限公司 一种色度色散测量方法、装置及数字相干接收机
CN105375978A (zh) * 2014-08-25 2016-03-02 深圳市中兴微电子技术有限公司 一种光传输网的色散检测方法和装置

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1970755A3 (en) * 1997-06-18 2014-08-27 Nippon Telegraph and Telephone Corporation White pulse source and applications
JP4401626B2 (ja) * 2002-07-05 2010-01-20 富士通株式会社 光信号を処理する方法及び装置
TWI360984B (en) * 2009-03-25 2012-03-21 Ind Tech Res Inst Method for receiving an optical ofdm signal and re
US8498542B2 (en) * 2010-01-21 2013-07-30 Ciena Corporation Multi-channel optical transceiver with offset quadrature amplitude modulation
US9264145B2 (en) * 2013-05-31 2016-02-16 Alcatel Lucent Optical receiver having a chromatic-dispersion compensation module with a multibranch filter-bank structure
CN104780037B (zh) * 2014-01-10 2019-04-30 深圳市中兴微电子技术有限公司 一种时钟恢复方法、装置及系统
JP6020696B1 (ja) * 2015-11-05 2016-11-02 Ｎｔｔエレクトロニクス株式会社 波長分散推定回路、光受信装置及び波長分散量推定方法
JP6376211B2 (ja) * 2016-11-30 2018-08-22 Ｎｔｔエレクトロニクス株式会社 波長分散補償装置、波長分散補償方法及び通信装置
CN108281877A (zh) * 2018-03-14 2018-07-13 成都师范学院 基于光谱角色散的啁啾激光脉冲频谱整形系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6204924B1 (en) * 1999-02-23 2001-03-20 Exfo Electro-Optical Engineering Inc. Method and apparatus for measuring polarization mode dispersion of optical devices
EP1318622A2 (en) * 2001-12-05 2003-06-11 Nortel Networks Limited Polarisation mode dispersion measurement and compensation
CN104579476A (zh) * 2013-10-22 2015-04-29 中兴通讯股份有限公司 光相干通信中色散估计方法及装置
CN104780131A (zh) * 2014-01-15 2015-07-15 深圳市中兴微电子技术有限公司 一种色度色散测量方法、装置及数字相干接收机
CN105375978A (zh) * 2014-08-25 2016-03-02 深圳市中兴微电子技术有限公司 一种光传输网的色散检测方法和装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3890211A4 *

Also Published As

Publication number Publication date
KR20210096191A (ko) 2021-08-04
JP7482128B2 (ja) 2024-05-13
EP3890211A1 (en) 2021-10-06
CN111262634A (zh) 2020-06-09
JP2022509301A (ja) 2022-01-20
EP3890211A4 (en) 2022-01-19
CN111262634B (zh) 2020-11-17

Similar Documents

Publication Publication Date Title
JP6405833B2 (ja) 信号処理装置及び信号処理方法
JP5146285B2 (ja) 周波数オフセット補償装置及び方法並びに光コヒーレント受信機
JP5545892B2 (ja) 信号生成回路および信号受信回路、信号生成回路、光信号送信装置、信号受信回路、光信号同期確立方法、および光信号同期システム
CN102461035B (zh) 用于对偏振分集复用信号进行盲解复用的方法和装置
US9698904B2 (en) Apparatus for monitoring optical signal to noise ratio, transmitter and communication system
CN109845144B (zh) 光接收机、光接收方法和光通信系统
EP3048746B1 (en) Method and device for estimation of chromatic dispersion in optical coherent communication
US8514922B2 (en) Filter coefficient control apparatus and method
US20160329970A1 (en) Clock recovery method, device and system and computer storage medium
US9397755B2 (en) Clock recovery method for ultra dense WDM systems
EP3884610B1 (en) Post-reception synchronization in a continuous variable quantum key distribution (cv-qkd) system
CN105612700B (zh) 用于表征光接收信号的色度色散的装置
EP2583424A1 (en) Method for phase and oscillator frequency estimation
WO2015106494A1 (zh) 一种色度色散测量方法、装置及数字相干接收机
JP2023506565A (ja) 周波数領域等化方法、等化器、光受信器、及びシステム
WO2020108632A1 (zh) 色散估计方法、装置、接收机及存储介质
CN106301593B (zh) 自适应盲偏振解复用处理方法和装置
US10505641B2 (en) Clock recovery for band-limited optical channels
JP5993042B2 (ja) 光伝送システム及び伝送路補償方法
Cheng et al. Training-aided joint frame and frequency synchronization for free space optical communication signals with low OSNR
CN110168967B (zh) 一种光接收机及延时估计方法
Ma et al. A novel high precision adaptive equalizer in digital coherent optical receivers
EP3133751A1 (en) Method for nonlinearity compensation in optical transmission systems
Yu et al. Basic Digital Signal Processing for Single-Carrier Signals
CN115733556A (zh) 光通信系统的相位噪声估计方法和相关装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19888798

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021531064

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 20217019935

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2019888798

Country of ref document: EP

Effective date: 20210630