JP5312384B2 - Digital sample processing method, digital sample processing apparatus, and program - Google Patents

Description

  The present invention relates to a digital sample processing method, a digital sample processing apparatus, and a program using a coherent technique.

  In recent years, a coherent transmission method using digital signal processing has been actively studied, and in particular, a coherent transmission method employing QPSK (Quadrature Phase Shift Keying) has been studied (for example, see Non-Patent Document 1). Compared with the conventional direct detection method, the reception sensitivity can be increased, and waveform processing such as chromatic dispersion and polarization mode dispersion can be accurately equalized by performing digital signal processing in the receiver. For this reason, especially in an optical device, it is difficult to compensate, and polarization mode dispersion, which has significantly limited the transmission distance of a high-speed optical signal, can be compensated, and the transmission distance of a 100 Gb / s / ch class optical signal is dramatic. Stretched.

  The digital coherent method employs a method in which quasi-static chromatic dispersion is compensated by a fixed digital filter, and fluctuation polarization mode dispersion is compensated by a small-tap adaptive filter based on a blind algorithm. As a result, long-distance transmission of an optical channel with high frequency and high frequency utilization efficiency has become possible. In addition to using multilevel modulation technology such as QPSK, polarization multiplexing technology is used to increase frequency utilization efficiency, and adaptation based on a blind algorithm typified by CMA (Constant modulus Algorithm) in the receiver Polarization separation is performed with a filter.

  In addition, in order to further improve the frequency utilization efficiency, research on a QAM (Quadrature Amplitude Modulation) format of 8 values or more has been conducted (for example, see Non-Patent Document 2), and the density of optical signals has been increased. These systems employ a decision feedback type least mean square (DD-LMS) algorithm as an adaptive equalization algorithm.

  The DD-LMS algorithm is an algorithm that can perform not only equalization of an optical reception signal but also phase recovery, but it is difficult to converge in a situation where a carrier frequency offset or a high-speed phase offset exists. .

  FIGS. 7A and 7B are block diagrams showing the configuration of a conventional receiver for optical multilevel polarization multiplexed signals. In FIG. 7A, the polarization multiplexed optical signal input to the optical front end via the optical transmission line is converted into a digital complex signal sequence at 2 samples / symbol in the A / D converters 1 to 4. The chromatic dispersion compensation circuit 5 compensates for the influence of chromatic dispersion received on the optical transmission line, and the butterfly adaptive filter 6 separates the components of the polarization multiplexed signal, equalizes the distortion due to polarization mode dispersion, the optical filter and the front The signal distortion factor at the end is equalized.

  At this time, each of the tap coefficient calculation circuits 7 and 8 performs initial convergence of tap coefficients for adaptive equalization using a CMA-based algorithm, and estimates a frequency offset and a phase offset by an algorithm such as a maximum likelihood estimation method. . Phase offset compensation circuits 9 and 10 perform phase offset compensation, frequency offset compensation circuits 11 and 12 perform frequency offset compensation, and finally, determination circuits 13 and 14 determine symbol points.

  After that, when the state becomes stable, as shown in FIG. 7B, in the tap coefficient calculation circuits 7a and 7b, the DD-LMS is used for the tap coefficient calculation, and the phase offset compensation circuits 9, 10 and the frequency offset compensation circuit are used. The algorithm for frequency offset compensation and phase offset compensation in 11 and 12 is also switched to a decision-oriented algorithm.

  When CMA is used for multi-level QAM signals, phase offset / frequency offset estimation is possible with the vitervi & vitervi algorithm or the maximum likelihood estimation method, but the algorithm has become complicated, and these estimation algorithms have moved to DD-LMS. Not used when. Moreover, CMA may supplement the same polarization component of the transmitted polarization multiplexed signal depending on the state of the received optical signal.

“10 x 111 Gbit / s, 50 GHz Spaced, POLMUX-RZ-DQPSK Transmission over 2375 km Employing Coherent Equalisation”, C.S.R.Flugder et al., OFC2007 paper: PDP22 ”112-Gb / s polarization-multiplexed 16-QAM on a 25-GHz WDM grid“, Peter J. Winzer et al., ECOC2008, Th3.E.5

  As described above, in the multi-value digital coherent reception system of eight or more values using the conventional blind algorithm for the initial convergence, when the polarization multiplexed signal is separated by polarization, the same polarization supplement problem or the tap coefficient calculation algorithm is used. After switching, it is necessary to switch the phase offset / frequency offset estimation algorithm, resulting in a problem that the circuit configuration becomes complicated. In the conventional algorithm, the signal to be separated from the X polarization and the Y polarization is separated into the X polarization, the X polarization, the Y polarization, and the Y polarization. There was a problem that it might be captured.

  The present invention has been made in consideration of such circumstances, and an object of the present invention is to simplify digital circuit processing that can simplify the circuit configuration and prevent the same polarization from being captured. A method, a digital sample processing apparatus, and a program are provided.

  In order to solve the above-described problem, the present invention provides a method for processing a digital sample sequence of a multilevel optical polarization multiplexed signal by a coherent receiver that receives the multilevel optical polarization multiplexed signal. Chromatic dispersion compensation step for compensating for chromatic dispersion of optical polarization multiplexed signal, and equalization for polarization separation, polarization mode dispersion compensation, and residual signal distortion compensation of the signal output from the chromatic dispersion compensation step A processing step, a frequency offset compensation step for compensating a frequency offset of the output signal obtained in the equalization processing step, a signal judgment step for performing symbol judgment of the output signal obtained in the frequency offset compensation step, and the frequency The frequency offset value used in the offset compensation step is the same as the output signal obtained in the frequency offset compensation step and the signal determination step. The frequency offset calculation step calculated using the output signal obtained in the step and the equalization coefficient used in the equalization processing step are calculated in the output signal obtained in the signal determination step and the frequency offset calculation step. A coefficient calculation step of calculating using the frequency offset value obtained and the output signal obtained in the equalization processing step, wherein the coefficient calculation step is a polarization component of the multilevel optical polarization multiplexed signal The equalization coefficient is calculated so that the Y polarization component is acquired using the X polarization component so that the Y polarization equalization coefficient does not correlate with the X polarization equalization coefficient. This is a digital sample processing method.

  According to the present invention, in the above invention, the coefficient calculation step calculates an error value of the equalization coefficient separately in a real part (Ich) and an imaginary part (Qch) of two polarization components of the digital sample sequence, The equalization coefficient is adaptively updated, and the equalization processing step restores the carrier phase of the output signal using the equalization coefficient calculated in the coefficient calculation step.

  The present invention is characterized in that, in the above invention, the frequency offset calculation step estimates the frequency offset value using only the sign of the output value of the signal determination step in the initial stage of convergence of the equalization coefficient calculation. To do.

  In order to solve the above-described problem, the present invention provides a coherent receiver that receives a multilevel optical polarization multiplexed signal, and a digital sample processing apparatus that processes a digital sample sequence of the multilevel optical polarization multiplexed signal. A chromatic dispersion compensation circuit that performs chromatic dispersion compensation processing of the multi-level optical polarization multiplexed signal, and polarization separation and polarization of the signal output from the chromatic dispersion compensation circuit based on an equalization coefficient given thereto. An equalization processing circuit that performs wave mode dispersion compensation and residual signal distortion compensation, a frequency offset compensation circuit that compensates a frequency offset of a signal output from the equalization processing circuit, and a signal output from the frequency offset compensation circuit A signal determination circuit that performs symbol determination of the frequency offset value used in the frequency offset compensation circuit, and an output signal of the frequency offset compensation circuit A frequency offset calculation circuit that calculates using an output signal of the signal determination circuit, and an equalization coefficient used in the equalization processing circuit, and a frequency offset calculated by the output signal of the signal determination circuit and the frequency offset calculation circuit. A coefficient calculation circuit that calculates a value and an output signal of the equalization processing circuit, and the coefficient calculation circuit equalizes the Y polarization equalization coefficient that is a polarization component of the multilevel optical polarization multiplexed signal The digital sample processing apparatus is characterized in that the equalization coefficient is calculated so as to obtain the Y polarization component using the X polarization component so that there is no correlation with the equalization coefficient of the X polarization.

  The present invention is the above invention, wherein the coefficient calculation circuit calculates an error value of the equalization coefficient separately in a real part (Ich) and an imaginary part (Qch) of two polarization components of a digital sample sequence, The equalization coefficient is adaptively updated, and the equalization processing circuit restores the carrier phase of the output signal using the equalization coefficient calculated by the coefficient calculation circuit.

  According to the present invention, in the above invention, the frequency offset calculation circuit calculates a frequency offset value by using a sign of an output value of the signal determination circuit in an initial stage of convergence of the equalization coefficient calculation by the coefficient calculation circuit. It is characterized by that.

  In order to solve the above-described problem, the present invention provides a coherent receiver that receives a multilevel optical polarization multiplexed signal, and a digital sample processing device that processes a digital sample sequence of the multilevel optical polarization multiplexed signal. A chromatic dispersion compensation function for performing chromatic dispersion compensation processing of the multilevel optical polarization multiplexed signal on a computer, polarization separation of the output signal obtained by the chromatic dispersion compensation function, polarization mode dispersion compensation, and residual signal distortion An equalization processing function for performing compensation, a frequency offset compensation function for compensating for a frequency offset of the output signal obtained by the equalization processing function, a signal determination function for performing symbol determination of the output signal obtained by the frequency offset compensation function, The frequency offset value used in the frequency offset compensation is the output signal obtained by the frequency offset compensation function and the output obtained by the signal determination function. Using the frequency offset calculation function, the output signal obtained by the signal determination function, the frequency offset value calculated by the frequency offset calculation function, and the output signal obtained by the equalization function Thus, the Y polarization component is used by using the X polarization component so that the equalization coefficient of the Y polarization, which is the polarization component of the multilevel optical polarization multiplexed signal, does not correlate with the equalization coefficient of the X polarization. As a result, a coefficient calculation function for calculating an equalization coefficient used in the equalization processing function is executed.

  According to the present invention, the circuit configuration can be simplified and the same polarization can be prevented from being captured.

It is a block diagram which shows the structural example of the optical coherent receiver by embodiment of this invention. It is a block diagram which shows the structural example of the loop filter by embodiment of this invention. 5 is a flowchart for explaining an operation of the optical coherent receiver according to the embodiment of the present invention. It is a figure which shows the post-demodulation constellation of the polarization multiplexing 16QAM signal by embodiment of this invention. It is a figure which shows the post-demodulation constellation of the polarization multiplexing 64QAM signal by embodiment of this invention. It is a figure which shows the transmission experiment result of the polarization multiplexing 16QAM signal by embodiment of this invention. It is a block diagram which shows the structure of the receiver of the conventional optical multilevel polarization multiplexed signal.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  The present invention is characterized in that initial convergence is performed using a blind algorithm (for example, MMA: Multi-modulus Algorithm) capable of reproducing a carrier phase in a tap coefficient calculation circuit of a butterfly FIR filter circuit. By using MMA, the phase offset is compensated at the output stage of the FIR filter, and the frequency offset calculation circuit continues to operate with the same algorithm even after switching to DD-LMS. Further, the result of the X polarization side tap coefficient calculation circuit is used in the Y polarization side tap coefficient calculation circuit. This produces an effect of preventing the same polarization from being captured. Details will be described below.

First, variables, constants, and symbols used in the description of this embodiment are shown.
X ₀ (t), Y ₀ (t): Output signal of wavelength difference compensation circuit X ₁ (t), Y ₁ (t): Output signal of butterfly FIR filter circuit X ₂ (t), Y ₂ (t ): output signal H _xx frequency offset compensation _{circuit, H xy, H yx, H} yy: tap coefficients of butterfly FIR filter mu: step size epsilon _x tap coefficients, epsilon _y: error in X polarization and Y polarization Signals θ _xn , θ _yn : Values for correcting the phase with respect to X polarization and Y polarization η: Real number coefficient that attenuates as time elapses ε _pol : Error signal ν: Forgetting factor α (n): X polarization output and Y Correlation values with polarization output S _x (n), S _y (n): output signals of determination circuit φ _x (n), φ _y (n): output signals of phase detection circuit φ _xn , φ _yn : NCO ( Input signals to Numerically-Controlled Oscillator) K1, K2: Proportional path in the phase detector circuit, the gain of the integral path R: constant corresponding to the multi-level number of the equalized signal

  FIG. 1 is a block diagram illustrating a configuration example of an optical coherent receiver according to an embodiment of the present invention. It should be noted that portions corresponding to those in FIG. The polarization multiplexed optical signal input to the optical front end via the optical transmission line is converted into a digital complex signal sequence at 2 samples / symbol in the A / D converters 1 to 4. The digital complex signal sequence is input to the chromatic dispersion compensation circuit 5 and compensates for the influence of chromatic dispersion received on the optical transmission line.

  More specifically, the chromatic dispersion compensation circuit 5 calculates, for example, a chromatic dispersion transfer function from the chromatic dispersion amount, and performs equalization in the frequency domain. Further, the impulse response may be calculated from the transfer function of chromatic dispersion, and the chromatic dispersion may be compensated by a time domain FIR filter. The digital complex signal sequence output from the chromatic dispersion compensation circuit 5 is input to the butterfly FIR filter circuit 6.

The polarization components output from the chromatic dispersion compensation circuit 5 are the signal X ₀ (t) and the signal Y ₀ (t), and the butterfly FIR filter components are H _xx , H _xy , H _yx , H _{When yy} is set, the signal X ₁ (t) and the signal Y ₁ (t) output from the butterfly FIR filter circuit 6 are given by the following equation (1).

Here, the tap coefficient _{H xx,} H _xy, H _yx, matrix H represented by _{H yy} is a matrix showing the FIR filter taps L, tap coefficient _{H xx,} the _{H xy,} tap coefficient computing circuit is set sequentially calculated tap coefficients at 20, the tap coefficients H _yx, the H _yy, sequentially calculated tap coefficients in the tap arithmetic circuit 21 is set. This butterfly FIR filter circuit 6 performs component separation of polarization multiplexed signals, distortion equalization due to polarization mode dispersion, and equalization of signal distortion factors at the optical filter and front end.

Next, the frequency offset compensation circuits 15 and 16 respectively use the digital offset signal X ₁ and Y output from the butterfly FIR filter circuit 6 using the frequency offset values calculated by the frequency offset calculation circuits 22 and 23, respectively. The frequency offset of ₁ is compensated, and finally the symbol points are determined by the determination circuits 13 and 14.

  Next, algorithm switching and frequency offset compensation of the tap coefficient calculation circuits 20 and 21 according to the present embodiment will be described. The tap coefficient calculation circuits 20 and 21 perform initial convergence using a blind algorithm capable of reproducing the carrier phase, for example, MMA (Multi-modulus Algorithm). Here, the tap coefficient calculation circuit 21 performs a calculation using the result of the tap coefficient calculation circuit 20 on the X polarization side to avoid the same polarization component supplement.

  By using MMA, the output signal of the butterfly FIR filter circuit 6 is automatically recovered in the carrier phase. Since the frequency offset between the light source of the optical transmitter and the local oscillation light source in the receiver remains, frequency offset compensation is performed at the output stage of the butterfly FIR filter circuit 6.

  After convergence of the tap coefficient and frequency offset estimation, the process proceeds to the DD-LMS algorithm. Since the tap coefficient calculated using MMA restores the symbol point of the multi-level signal with respect to the input signal sequence, the equalization coefficient can be inherited as it is when the algorithm is switched to DD-LMS.

  When shifting to DD-LMS, it is not necessary to switch the algorithm for frequency offset estimation, and the frequency offset calculation circuits 22 and 23 continue to operate and compensate for the frequency offset in the same way as during MMA operation.

The update formula used in the tap coefficient calculation circuits 20 and 21 is given by formula (2). n represents the digital sample sequence to be processed. For example, when the number of taps of the matrix H is 5, the complex conjugate X ₀ ^* (n) of the output of the chromatic dispersion compensation circuit 5 is a complex signal sequence vector of length 5. θ _xn and θ _yn are phase correction terms estimated by the frequency offset calculation circuits 22 and 23 for the X polarization component and the Y polarization component, respectively. μ is a coefficient update step size. ε _x and ε _y represent error signals. As in the conventional method, reception is possible without separately providing the phase offset compensation circuits 9 and 10 shown in FIGS.

The signals X ₂ and Y ₂ input to the respective polarization component determination circuits 13 and 14 are given by Expression (3).

In the present embodiment, since the same update formula is used regardless of the algorithm, only the calculation formulas of the error signals ε _x and ε _y and the phase correction terms θ _xn and θ _yn calculated by the frequency offset calculation circuits 22 and 23 are obtained. Switching between initial convergence and after convergence may be performed as described below. The error signal when MMA is used is calculated using Equation (4). Re [ ^* ] is a function that outputs the real part of the variable, and Im [ ^* ] is a function that outputs the imaginary part of the variable. R is a constant, and the optimum value varies depending on the number of multilevel signals to be equalized.

Here, the feature of the present embodiment is that the error signal ε _pol expressed by Expression (5) using the output signal X ₂ of X polarization is added to the error signal of Y polarization, thereby supplementing the same polarization. There is a place to avoid. α (n) represents the correlation between the X polarization output and the Y polarization output in sample n. η is a real coefficient that attenuates over time. ν is a real forgetting factor smaller than 1.

In Equation (5), supplementation of the same polarization is avoided by using one symbol, but the error signal ε _pol may be calculated by extending α (n) and using X ₂ of several symbols before and after. After the tap coefficient is converged by MMA, the tap coefficient is updated by DD-LMS according to Equation (6). S _x (n) and S _y (n) are signals indicating the output results of the determination circuits 13 and 14 for the nth symbol.

  Next, the frequency offset calculation circuits 22 and 23 will be described. The frequency offset calculation circuits 22 and 23 include phase detection circuits 22-1 and 23-1, loop filters 22-2 and 23-2, and NCO (Numerically-Controlled Oscillator) 22-3 and 23-3, respectively. Yes.

FIG. 2 is a block diagram illustrating a configuration example of the loop filters 22-2 and 23-2 of the frequency offset calculation circuits 22 and 23 according to the present embodiment. When a general active filter as shown in FIG. 2 is used as the loop filters 22-2 and 23-2, signals φ _xn and φ _yn input to NCO (Numerically-Controlled Oscillator) 22-3 and 23-3 are , Given by equation (7).

φ _x and φ _y are outputs of the phase detection circuits 22-1 and 23-1. K ₁ and K ₂ are the gains of the active filter, K ₁ is the proportional path, and K ₂ is the gain of the integration path. By adjusting K ₁ and K ₂ , a frequency offset amount that can be compensated and a convergence time can be set. The output phases θ _xn and θ _yn of the NCOs 22-3 and 23-3 are obtained by time integration of φ _xn and φ _yn , respectively.

In the present embodiment, in order to increase the convergence speed, the phase detection circuits 22-1 and 23-1 switch the phase detection method depending on the convergence state of the frequency offset estimation value. When the estimated frequency offset is far from the true value, the reliability of the outputs of the determination circuits 13 and 14 is low. At this time, the phase is detected according to the equation (8) using the signs of the outputs of the determination circuits 13 and 14. Sign ( ^* ) is a function that outputs the sign of a variable.

  When the frequency offset value and the coefficient of the equalization filter converge to some extent and the reliability of the outputs of the determination circuits 13 and 14 is improved, the phase detection circuits 22-1 and 23-1 accurately follow the equation (9). A phase error is detected and the accuracy of frequency offset estimation is improved.

  FIG. 3 is a flowchart for explaining the operation of the optical coherent receiver according to the embodiment of the present invention. First, the chromatic dispersion compensation circuit 5 performs chromatic dispersion compensation processing of the optical signal converted into a digital complex signal sequence at 2 samples / symbol by the A / D converters 1 to 4 (step S1). Next, the butterfly FIR filter circuit 6 performs polarization separation, polarization mode dispersion compensation, and residual signal distortion compensation of the signal output from the chromatic dispersion compensation circuit 5 according to a predetermined equalization coefficient (equalization processing). (Step S2).

  Next, the frequency offset compensation circuits 15 and 16 compensate for the frequency offset of the equalized signal (step S3), and the determination circuits 13 and 14 perform symbol determination of the frequency offset compensated signal (step S3). S4). Next, the frequency offset calculation circuits 22 and 23 use the frequency offset values used in the frequency offset compensation by using the output signals of the frequency offset compensation circuits 15 and 16 and the output signals (symbol determination results) of the determination circuits 13 and 14. Calculation is performed (step S5).

  The tap coefficient calculation circuits 20 and 21 output the output signals (symbol determination results) of the determination circuits 13 and 14, the frequency offset value estimated by the frequency offset calculation circuits 22 and 23, and the output signal of the butterfly FIR filter circuit 6. Is used to calculate the equalization coefficient used in the equalization process (step S6). At this time, the tap coefficient calculation circuit 21 uses the output of the tap coefficient calculation circuit 20 on the X polarization side so that the equalization coefficient of the Y polarization does not correlate with the equalization coefficient of the X polarization. The equalization coefficient is calculated so as to obtain the Y polarization component which is a different polarization component of the wave multiplexed signal. This avoids the same polarization component supplement.

  4 and 5 are diagrams showing constellations obtained by demodulating a 160 Gb / s, 16 QAM signal and a polarization multiplexed signal of 240 Gb / s, 64 QAM signal according to the present embodiment. FIG. 6 is a diagram showing a result of a transmission experiment using a polarization-division multiplexed 16QAM signal of 160 Gb / s according to the present embodiment. The error correction code (FEC) threshold value of 9.1 dB is exceeded after 3120 km transmission, and it can be seen that a multilevel signal can be received with good quality even after long distance transmission.

  As described above, according to the present invention, in a coherent optical transmission system, a multilevel optical polarization multiplexed signal can be received with high quality without supplementing the same polarization without using a known signal.

1-4 A / D
DESCRIPTION OF SYMBOLS 5 Chromatic dispersion compensation circuit 6 Butterfly type FIR filter circuit 13, 14 Judgment circuit 15, 16 Frequency offset compensation circuit 20, 21 Tap coefficient calculation circuit 22, 23 Frequency offset calculation circuit 22-1, 23-1 Phase detection circuit 22-2 , 23-2 Loop filter 22-3, 23-3 NCO

Claims (7)

A method of processing a digital sample sequence of the multilevel optical polarization multiplexed signal with a coherent receiver that receives the multilevel optical polarization multiplexed signal,
A chromatic dispersion compensation step for performing chromatic dispersion compensation processing of the multilevel optical polarization multiplexed signal;
Equalization processing step for performing polarization separation, polarization mode dispersion compensation, and residual signal distortion compensation of the signal output from the chromatic dispersion compensation step;
A frequency offset compensation step for compensating a frequency offset of the output signal obtained in the equalization processing step;
A signal determination step for performing symbol determination of the output signal obtained in the frequency offset compensation step;
A frequency offset calculation step for calculating the frequency offset value used in the frequency offset compensation step using the output signal obtained in the frequency offset compensation step and the output signal obtained in the signal determination step;
The equalization coefficient used in the equalization processing step uses the output signal obtained in the signal determination step, the frequency offset value calculated in the frequency offset calculation step, and the output signal obtained in the equalization processing step. The coefficient calculation step includes the following steps:
The Y polarization component is acquired using the X polarization component so that the Y polarization equalization coefficient that is the polarization component of the multi-level optical polarization multiplexed signal does not correlate with the X polarization equalization coefficient. A digital sample processing method characterized in that an equalization coefficient is calculated as follows. The coefficient calculation step includes:
An error value of the equalization coefficient is separately calculated in the real part (Ich) and the imaginary part (Qch) of the two polarization components of the digital sample sequence, and the equalization coefficient is updated adaptively.
The equalization processing step includes
The digital sample processing method according to claim 1, wherein the carrier phase of the output signal is restored using the equalization coefficient calculated in the coefficient calculation step. The frequency offset calculating step includes:
3. The digital sample processing method according to claim 2, wherein in the initial stage of convergence of the equalization coefficient calculation, the frequency offset value is estimated using only the sign of the output value of the signal determination step. A coherent receiver for receiving a multilevel optical polarization multiplexed signal, a digital sample processing device for processing a digital sample sequence of the multilevel optical polarization multiplexed signal,
A chromatic dispersion compensation circuit that performs chromatic dispersion compensation processing of the multilevel optical polarization multiplexed signal;
An equalization processing circuit for performing polarization separation, polarization mode dispersion compensation, and residual signal distortion compensation of a signal output from the chromatic dispersion compensation circuit based on a given equalization coefficient;
A frequency offset compensation circuit for compensating for a frequency offset of the signal output from the equalization processing circuit;
A signal determination circuit for performing symbol determination of the signal output from the frequency offset compensation circuit;
A frequency offset calculation circuit that calculates a frequency offset value used in the frequency offset compensation circuit using an output signal of the frequency offset compensation circuit and an output signal of the signal determination circuit;
A coefficient calculation circuit that calculates an equalization coefficient used in the equalization processing circuit by using an output signal of the signal determination circuit, a frequency offset value calculated by the frequency offset calculation circuit, and an output signal of the equalization processing circuit And
The coefficient arithmetic circuit is
The Y polarization component is acquired using the X polarization component so that the Y polarization equalization coefficient that is the polarization component of the multi-level optical polarization multiplexed signal does not correlate with the X polarization equalization coefficient. A digital sample processing device, characterized in that an equalization coefficient is calculated. The coefficient arithmetic circuit is
An error value of the equalization coefficient is separately calculated in the real part (Ich) and the imaginary part (Qch) of the two polarization components of the digital sample sequence, and the equalization coefficient is updated adaptively.
The equalization processing circuit includes:
The digital sample processing apparatus according to claim 4, wherein the carrier phase of the output signal is restored using the equalization coefficient calculated by the coefficient arithmetic circuit. The frequency offset calculation circuit includes:
6. The digital sample processing device according to claim 5, wherein in an initial stage of convergence of equalization coefficient calculation by the coefficient calculation circuit, a frequency offset value is calculated using a sign of an output value of the signal determination circuit. In a coherent receiver that receives a multilevel optical polarization multiplexed signal, a computer of a digital sample processing device that processes a digital sample sequence of the multilevel optical polarization multiplexed signal,
A chromatic dispersion compensation function for performing chromatic dispersion compensation processing of the multilevel optical polarization multiplexed signal;
An equalization processing function for performing polarization separation of the output signal obtained by the chromatic dispersion compensation function, polarization mode dispersion compensation, and residual signal distortion compensation;
A frequency offset compensation function for compensating a frequency offset of the output signal obtained by the equalization processing function;
A signal determination function for performing symbol determination of the output signal obtained by the frequency offset compensation function;
A frequency offset calculation function for calculating a frequency offset value used in the frequency offset compensation using an output signal obtained by a frequency offset compensation function and an output signal obtained by the signal determination function;
Using the output signal obtained by the signal determination function, the frequency offset value calculated by the frequency offset calculation function, and the output signal obtained by the equalization processing function, the deviation of the multilevel optical polarization multiplexed signal is obtained. The equalization processing function acquires the Y polarization component using the X polarization component so that the equalization coefficient of the Y polarization that is the wave component does not correlate with the equalization coefficient of the X polarization. A program for executing a coefficient calculation function for calculating an equalization coefficient used in.

Images