JP2010226254A - Digital signal processing circuit and optical receiving device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a digital signal processing circuit capable of achieving non-linear compensation with higher accuracy, and to provide an optical receiving device. <P>SOLUTION: A wavelength dispersion compensated unit, based on a fixed FIR filter is divided into a plurality of sections which are respectively constituted of wavelength dispersion compensation parts 4-1-1 to 4-1-N; phase rotation parts 4-13-1 to 4-13-N, wavelength dispersion compensation parts 4-2-1 to 4-2-N; and phase rotation parts 4-14-1 to 4-14-N, and each of non-linear compensated parts which are respectively constituted of square operation parts 4-3-1 to 4-3-N, 4-4-1 to 4-4-N is inserted between the sections. Furthermore, each of addition parts 4-9-1, 4-10-1 uses both of its own polarization receiving data and the other polarization receiving data to achieve more accurate compensation. In order to fully develop the compensation effect further, skews between received X polarization data and received Y polarization data in the receiver and the gains in the phase rotation parts 4-13, 1, 4-14-1, etc. are controlled. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a digital signal processing circuit and an optical receiver.

  In recent years, in an optical transmission system, it is desired to increase the transmission capacity, and it is necessary to increase the transmission capacity while maintaining the transmission distance of the system. However, there is a trade-off relationship between transmission capacity and transmission distance, and this trade-off is caused by nonlinear distortion that occurs in the transmission line fiber.

  In a conventional optical transmission system, the trade-off relationship is overcome by using a technique such as a method of compensating chromatic dispersion for each optical amplification repeater, a Raman amplification technique, a low-loss fiber, and the like. However, the problem that the compensation accuracy is limited by the minimum unit of the compensation amount of the dispersion compensating fiber and the transmission distance and the frequency use efficiency are limited by nonlinearity has become remarkable. In addition, there is a problem that uncompensated chromatic dispersion accumulates depending on the signal wavelength due to the chromatic dispersion slope.

  Therefore, as a technique for relaxing the trade-off relationship between transmission capacity and transmission distance, nonlinear waveform distortion suppression / compensation technology has been researched and developed. As a cause of nonlinear waveform distortion in a channel, there is self phase modulation (SPM) in which the phase of light rotates in proportion to the signal amplitude intensity of its own channel. In the coherent reception method, since the optical phase of the signal light is also detected, SPM waveform distortion can be compensated in principle by signal processing at the receiving end. So far, SPM waveform distortion compensation has been studied, but has been studied (for example, see Non-Patent Documents 1 and 2).

K. Kikuchi, M. Fukase, and S. Kim, “Electronic post-compensation for nonlinear phase noise in a 1000-km 20-Gbit / s optical QPSK transmission system using the homodyne receiver with digital signal processing,” OFC2007, paper OTuA2 , 2007. G. Charlet, N. Maaref, J. Renaudier, H. Mardoyan, P. Tran, S. Bigo, “Transmission of 40Gb / s QPSK with coherent detection over ultra-long distance improved by nonlinearity mitigation,” ECOC2006, paper Th4. 3.4, 2006.

  By the way, the waveform distortion due to non-linearity changes depending on the time waveform of the optical electric field intensity of the signal light. However, since the time waveform of the optical electric field strength of the signal light changes in the transmission line due to chromatic dispersion in the optical fiber transmission line, the time waveform of the signal light detected at the receiving end is actually non-linear in the transmission line. It differs from the time waveform at the time of waveform distortion. This is a factor that limits the compensation effect and application range.

  In the prior art described above, a transmission path is constructed that returns the chromatic dispersion value to near zero for each optical amplification relay. However, in an actual transmission line, the chromatic dispersion is rarely reset for each optical repeater, and the time waveform of the signal light at the time of input to the optical fiber transmission line after the optical amplification repeater is different for each span.

  For example, when compensation is performed using a chromatic dispersion compensating fiber, uncompensated chromatic dispersion is accumulated because compensation accuracy is limited by the minimum unit of compensation amount. In general, since the amount of chromatic dispersion varies depending on the wavelength due to the chromatic dispersion slope, the compensation accuracy is still limited, and uncompensated chromatic dispersion accumulates.

  Thus, in the prior art, there is a limit to the accuracy of optical chromatic dispersion compensation, and uncompensated chromatic dispersion accumulates. Further, in optical chromatic dispersion compensation using a chromatic dispersion compensating fiber, there is a problem that dispersion design is complicated and the cost of supplies of the dispersion compensating fiber is high.

  In recent years, a method for electrically compensating for chromatic dispersion collectively at the receiving end using digital signal processing or the like has been researched and developed. In this case, since there is no optical dispersion compensation, the chromatic dispersion amount is monotonously accumulated. Accordingly, since the waveform of the signal light input to each transmission span is different for each span, the waveform distortion due to the nonlinear effect is also different for each span, and there is a problem that nonlinear distortion compensation cannot be applied.

  In the calculation of nonlinear waveform distortion compensation, it is necessary to change the compensation parameter depending on the magnitude of distortion. The magnitude of the nonlinear waveform distortion varies depending on the transmission conditions such as the optical power of the signal light, the effective cross-sectional area of the transmission line, the material of the optical fiber, and the propagation direction profile of the optical power by Raman amplification. Among these, the dependence on transmission power is particularly large. Therefore, a mechanism for adjusting the compensation parameter according to the transmission condition is necessary.

  In general, a coherent receiver mixes with local light and detects the amplitude / phase waveform of signal light using local light as reference light. In this case, four time waveforms are detected in order to separate the X-polarized wave and the Y-polarized wave, and to detect the in-phase component and the quadrature component, respectively. Since mutual time delays (skew) that occur in these four detection processes occur, in general coherent reception, the subsequent digital signal processing circuit has a function to compensate the residual skew error, so the accuracy of skew adjustment is improved. rough. Therefore, there is a problem in that it is insufficient for applying nonlinear waveform distortion compensation, and a mechanism for adjusting skew with high accuracy is required.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a digital signal processing circuit and an optical receiver capable of realizing nonlinear compensation with higher accuracy.

  In order to solve the above-described problem, the present invention is a digital signal processing circuit in an optical fiber transmission system, which is composed of n (n is a natural number) block compensation arithmetic circuit, and the first compensation arithmetic circuit is a phase compensation circuit. A first chromatic dispersion compensation unit that receives a digital received signal including information and compensates for chromatic dispersion; a first square computation unit that computes the square of the electric field amplitude from the output of the first chromatic dispersion compensation unit; A first amplifying unit that multiplies an output from the first square operation unit by a nonlinear coefficient and a nonlinear effective fiber length of an optical fiber, and an output signal from the first chromatic dispersion compensation unit. And a first phase rotation unit that performs phase rotation based on output information from the first amplification unit, and the k-th (k = 2 to n natural number) compensation arithmetic circuit has the (k−1) th Wavelength relative to the output from the phase rotation part A k-th chromatic dispersion compensation unit that compensates for dispersion, a k-th square computation unit that computes the square of the electric field amplitude from the output of the k-th chromatic dispersion compensation unit, a nonlinear coefficient and a nonlinear effective fiber length of the optical fiber, Is multiplied by the output from the k-th square computing unit, and the output signal from the k-th chromatic dispersion compensation unit is based on the output information from the k-th amplification unit. A digital signal processing circuit comprising a k-th phase rotation unit that performs phase rotation.

  The present invention is the digital signal processing circuit in the optical fiber transmission system according to the above invention, comprising a compensation arithmetic circuit of n (n is a natural number) block, the first compensation arithmetic circuit is polarization-separated, A first chromatic dispersion compensation unit that compensates chromatic dispersion for each polarization with respect to a digital reception signal including phase information, and a square of the electric field amplitude of each polarization from the output of the first chromatic dispersion compensation unit. A first square computing unit for computing, a first amplifying unit for multiplying the output from the first square computing unit for each polarization by the nonlinear coefficient and nonlinear effective fiber length of the fiber for each polarization, and the other A first multiplier that multiplies the nonlinear coefficient of the non-polarized fiber and the nonlinear effective fiber length, and adds the output of the first amplifier and the output of the first multiplier of the other polarization. A first two-input adder and the first chromatic dispersion; A first phase rotation unit that performs phase rotation on the output signal from the compensation unit based on output information from the first two-input addition unit, and kth (k is from 2 to n) The (natural number) compensation calculation circuit includes a kth chromatic dispersion compensation unit that compensates chromatic dispersion for each polarization with respect to an output from the (k−1) th phase rotation unit of each polarization, a k-th square computing unit that computes the square of the electric field amplitude of each polarization from the output of the k chromatic dispersion compensation unit, and the nonlinear coefficient and nonlinear effective fiber length of the fiber of each polarization. A k-th amplification unit for multiplying the output from the square calculation unit, a k-th multiplication unit for multiplying the nonlinear coefficient of the other polarization fiber and the nonlinear effective fiber length, and an output of the k-th amplification unit And the k-th two-input adder for adding the output of the k-th multiplier of the other polarization, and the k-th wavelength component A digital signal processing circuit comprising: a k-th phase rotation unit that performs phase rotation on the output signal from the compensation unit based on output information from the k-th two-input addition unit. is there.

  According to the present invention, in the above invention, for each of the n-block compensation arithmetic circuits, a (two or more integers) FIR filters connected in cascade and the outputs of the a F FIR filters are squared. Further comprising an all-input adder that adds all of the outputs from the first square calculator or the outputs of the k-th square calculator and the a square calculator. The a FIR filters perform a convolution operation on a first FIR filter that performs a convolution operation on the output of the k-th chromatic dispersion compensation unit and an output of a previous stage FIR filter connected in cascade ( a-1) second to a-th FIR filters, and the k-th phase rotation unit outputs an output signal from the k-th chromatic dispersion compensation unit from the all-input addition unit. Perform phase rotation based on output information It is characterized in.

  According to the present invention, in the above invention, a convolution operation is performed on each polarization of the output of the Lth (L = 1 to n) chromatic dispersion compensator for each of the n block compensation operation circuits. An L-th FIR filter, a square calculation unit that calculates the square of the electric field amplitude of each polarization from the output of the L-th FIR filter, and a nonlinear coefficient and a nonlinear effective fiber length of each polarization fiber. An amplifying unit for multiplying the output from the square computing unit, a multiplying unit for multiplying the nonlinear coefficient of the other polarization fiber and a nonlinear effective fiber length, and the multiplication of the output of the amplifying unit and the other polarization A three-input adder for adding the output of the second polarization unit and the output of the L-th multiplier of the other polarization, and the L-th two-input adder is connected to the output of the L-th amplifier Adding the output of the 3-input adder, and the L-th phase rotation unit , Based on output information from the L-th two-input adder, with respect to the digitally received signal that is phase-separated and includes phase information, or an output signal from the (L−1) -th chromatic dispersion compensator. And performing phase rotation.

  According to the present invention, in the above invention, the first to kth chromatic dispersion compensation units include a minority tap filter with reduced taps, and the output signal of the Mth (= n) phase rotation unit is predetermined. And a multi-tap filter with an increased number of taps, which compensates for errors in the amount of chromatic dispersion compensation by the minority tap filter.

  According to the present invention, in the above invention, the first to kth chromatic dispersion compensation units include a minority tap filter with reduced taps, and the output signal of the Mth (= n) phase rotation unit is predetermined. And a multi-tap filter with an increased number of taps, which compensates for errors in the amount of chromatic dispersion compensation by the minority tap filter.

  The present invention is characterized in that, in the above-mentioned invention, a gain control unit for controlling a gain of an input signal to the first to k-th phase rotation units based on a predetermined signal light power.

  The present invention is the above invention, wherein the dispersion compensation amount of the kth chromatic dispersion compensator is set to 1, 2,..., K in order from the receiving end in a transmission section sandwiched between two optical amplifiers. The chromatic dispersion amount in the k-th section is set.

  In order to solve the above-described problem, the present invention is an optical receiver that includes the digital signal processing circuit and receives an optical signal composed of X-polarized light and Y-polarized light. A skew adjustment unit that adjusts the skews of the in-phase and quadrature phase components of the polarization, a quality monitor unit that monitors the signal quality after demodulation of the optical signal, and signal quality information that is output from the quality monitor unit And a control unit for controlling the skew amount of the skew adjusting unit.

  According to the present invention, in the above invention, the quality monitoring unit includes a waveform distortion estimation unit that estimates a change in the optical electric field waveform due to self-phase modulation from the received signal, and an optical electric field waveform from the demodulated signal of the optical signal. A waveform distortion detection unit for detecting a waveform distortion amount, wherein the control unit changes the optical electric field waveform estimated by the waveform distortion estimation unit and the waveform distortion amount of the optical electric field waveform detected by the waveform distortion detection unit. The skew amount of the skew adjustment unit is controlled on the basis of the difference between the skew adjustment unit and the skew adjustment unit.

  According to the present invention, in the above invention, the quality monitoring unit includes a waveform distortion estimation unit that estimates a change in the optical electric field waveform due to self-phase modulation from the received signal, and an optical electric field waveform from the demodulated signal of the optical signal. A waveform distortion detection unit for detecting a waveform distortion amount, wherein the control unit changes the optical electric field waveform estimated by the waveform distortion estimation unit and the waveform distortion amount of the optical electric field waveform detected by the waveform distortion detection unit. The gain determining the amount of phase rotation of the in-phase component and the quadrature component of the X-polarized wave and the Y-polarized component is controlled on the basis of the difference between the gains.

  According to the present invention, it is possible to compensate for the nonlinear effect based on the amount of chromatic dispersion compensation, and it is possible to obtain an advantage that highly accurate nonlinear compensation can be performed by repeating a plurality of times.

It is a block diagram which shows the structure of the digital signal processing circuit by 1st Embodiment of this invention. It is a block diagram which shows the structure of the digital signal processing circuit by this 2nd Embodiment. In the digital signal processing circuit according to the third embodiment, when SPM compensation is performed, the signal quality is monitored to finely adjust the skew adjustment unit. It is a conceptual diagram for demonstrating an example of the constellation figure in quaternary phase modulation (QPSK). It is a conceptual diagram which shows an example of the result of having evaluated the Q value by changing the skew between X polarization and Y polarization. It is a block diagram which shows the structure of the dispersion compensation & SPM compensation circuit of the digital signal processing circuit by this 4th Embodiment. It is a block diagram which shows the structure of the digital signal processing circuit by this 5th Embodiment. It is a block diagram which shows the structure of the digital signal processing circuit by this 6th Embodiment. In this 6th Embodiment, it is a conceptual diagram which shows an example of the result of having changed the gain and evaluating Q value. It is a block diagram which shows the structure of the digital signal processing circuit by this 7th Embodiment. It is a conceptual diagram which shows the shift | offset | difference between the target point and the measurement point in distribution of the constellation for demonstrating the detection method of a nonlinear distortion amount. It is a block diagram which shows the structure of the gain control parts 17-1 to 17-N of the digital signal processing circuit by this 7th Embodiment. It is a block diagram which shows the structure which estimates only SPM phase rotation as a waveform distortion. It is a block diagram which shows the structure of the digital signal processing circuit by this 8th Embodiment. It is a block diagram which shows the structure of the digital signal processing circuit by this 9th Embodiment. It is a block diagram which shows the structure of the digital signal processing circuit by this 10th Embodiment. It is a block diagram which shows the structure of the transmitter at the time of using this invention for reception of orthogonal frequency division multiplexing (OFDM) signal light. It is a conceptual diagram which shows the effect at the time of using for reception of orthogonal frequency division multiplexing (OFDM) signal light. It is a block diagram which shows the structure of the digital signal processing circuit by the modification of this invention. It is a block diagram which shows the structure of the other digital signal processing circuit by the modification of this invention. It is a block diagram which shows the structure of the minority tap filter by this modification. It is a block diagram which shows the structure of the multi-tap filter by this modification.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
In general, a coherent reception method is generally a method of performing polarization separation using an adaptive FIR filter after chromatic dispersion compensation using a fixed FIR (Finite Impulse Response) filter. In the present invention, a chromatic dispersion compensation unit using a fixed FIR filter is divided into a plurality of sections, and a non-linear compensation unit is inserted between the sections, and a compensation calculation is performed using the signal light power in the transmission line fiber as input information. I do. In addition, the non-linear compensation of the X polarization includes a control mechanism for compensating more accurately by using both the reception data of the X polarization itself and the reception data of the Y polarization. Furthermore, in order to fully exhibit this compensation effect, it has a skew adjustment unit for sufficiently compensating the skew in the receiver between the received X polarization data and the received Y polarization data, and its control unit. It is characterized by.

A. First Embodiment First, a first embodiment of the present invention will be described.
FIG. 1 is a block diagram showing a configuration of a digital signal processing circuit according to the first embodiment of the present invention. In the figure, ADCs (Analog Digital Converters) 1-1 to 1-4 convert an X polarization Re component, an X polarization Im component, a Y polarization Re component, and a Y polarization Im component into digital signals. Convert to The skew adjustment units 2-1 to 2-4 are different depending on four paths, and the delay time from the time when the light is input to the 90 ° hybrid from the output end of the optical fiber to the time when the light is input to the ADC 1-1 to 1-4 To compensate. The complex synthesis circuit 3-1 complex-synthesizes the in-phase component waveform and the quadrature-phase component waveform of the X polarization, and the complex synthesis circuit 3-2 performs the in-phase component waveform and the quadrature-phase component waveform of the Y-polarization. And complex. The dispersion compensation & SPM compensation circuit 4 performs dispersion compensation and SPM compensation for the X polarization and the Y polarization. The polarization separation circuit 5 performs polarization separation on the digital signal subjected to dispersion compensation and SPM compensation using an adaptive FIR filter.

  The dispersion compensation & SPM compensation circuit 4 includes N chromatic dispersion compensation units 4-1-1-4-1-4 -N, 4-2-1-4-2 -N for each of the X polarization and the Y polarization. Square arithmetic units 4-3-1 to 4-3-N, 4-4-1 to 4-4-N, first amplifying units 4-5-1 to 4-5-N, 4-6-1 to 4-6-N, multipliers 4-7-1 to 4-7-N, 4-8-1 to 4-8-N, adders 49-1 to 4-9-N, 4-10- 1 to 4-10-N, and phase rotation units 4-13-1 to 4-13-N and 4-14-1 to 4-14-N.

  The chromatic dispersion compensators 4-1-1, 4-2-1 each receive polarization-separated digital signals including phase information and compensate the chromatic dispersion for each polarization. The chromatic dispersion compensators 4-1 -k, 4-2 -k (k = 2, 3,..., N) receive the outputs from the phase rotation units 4-14-(k−1) of the respective polarizations as inputs. The chromatic dispersion is compensated for each polarization.

  The square calculation units 4-3-1 to 4-3-N and 4-4-1 to 4-4-N are respectively connected to the preceding chromatic dispersion compensation units 4-1-1 to 4-1-N, 4- The square of the electric field amplitude of each polarization from 2-1 to 4-2N is calculated. The first amplifying units 4-5-1 to 4-5-N and 4-6-1 to 4-6-N respectively calculate the nonlinear coefficient and the nonlinear effective fiber length of each polarization fiber. Multiplies the output from the square wave computing units 4-3-1 to 4-3-N, 4-4-1 to 4-4-N. Each of the multipliers 4-7-1 to 4-7 -N and 4-8-1 to 4-8 -N multiplies the nonlinear coefficient of the other polarization fiber by the nonlinear execution fiber length.

  The adders 4-9-1 to 4-9-N and 4-10-1 to 4-10-N respectively include first amplifiers 4-5-1 to 4-5-N and 4-6. The outputs of 1-4-6-N and the outputs of the multipliers 4-7-1 to 4-7-N and 4-8-1 to 4-8-N corresponding to the other polarization are added. The phase rotators 4-13-1 to 4-13-N and 4-14-1 to 4-14-N are respectively chromatic dispersion compensating units 4-1-1 to 4-1-N, 4- Phase rotation is performed on the outputs from 2-1 to 4-N based on the information from the adder.

  In general, in the coherent reception method, local light and signal light are mixed using an optical 90 ° hybrid, and the time waveform of the amplitude and phase of the signal light is detected using local light as a reference. There, the optical electric field of the signal light is detected by separating it into a waveform having the same phase component as that of the local light and a waveform having a quadrature phase component.

  In the case of coherent reception with a polarization diversity configuration, signal time waveforms of the same phase component and quadrature phase component are detected for the X polarization and the Y polarization, respectively. For detection, a photoelectric converter such as a balanced photodiode is used and converted into an electrical analog signal. The time waveform is converted into a digital signal by the ADCs 1-1 to 1-4, and the subsequent signal processing is performed using digital signal processing. In the digital signal processing, the delay time from the time when the light is input to the 90 ° hybrid from the output end of the optical fiber to the time when the light is input to the ADCs 1-1 to 1-4 is the same phase component of the X polarization component, orthogonal. Since it differs depending on the four paths of the phase component, the in-phase component of Y-polarized wave, and the quadrature phase component, the skew adjusting units 2-1 to 2-4 for compensating for this are required.

  For example, the complex synthesis circuits 3-1 and 3-2 mathematically synthesize the time waveform of the optical signal light including the carrier phase by complex-combining the waveform of the in-phase component and the quadrature component of the X polarization. It becomes possible to express. Waveform distortion due to transmission line nonlinearity is expressed as an effect of rotating the signal light phase in proportion to the square of the optical electric field amplitude, as shown in the following equation.

  Therefore, in order to compensate for this SPM waveform rotation, a phase rotation amount proportional to the square of the optical electric field amplitude may be given in the reverse direction. However, since the temporal waveform of the optical electric field intensity changes due to chromatic dispersion, the waveform distortion due to SPM and chromatic dispersion is intertwined in a complicated manner. In order to achieve this approximately, as shown in FIG. 1, the chromatic dispersion compensation units 4-1-1 to 4-1-N and 4-2-1 to 4-2-2N are divided into a plurality of sections. SPM compensation units (square computing units 4-3-1 to 4-3-N, 4-4-1 to 4-4-N, and first amplification unit 4-5 between the divided chromatic dispersion compensation sections. 1-4-5-N, 4-6-1 to 4-6-N, multiplying units 4-7-1 to 4-7-N, 4-8-1 to 4-8-N, adding unit 4- 9-1 to 4-9-N, 4-10-1 to 4-10-N, and phase rotation units 4-13-1 to 4-13-N, 4-14-1 to 4-14-N) Insert.

  For example, it is assumed that the chromatic dispersion compensation amount is divided for each optical fiber transmission section (span) sandwiched between optical amplifiers. In the first chromatic dispersion compensation section composed of the chromatic dispersion compensation units 4-1-1, 4-2-1, by inputting the optical electric field waveform detected at the receiving end and compensating the chromatic dispersion of the final span, The time waveform of the signal light at the input end of the final span can be reproduced. Next, this time waveform is obtained from the first SPM compensation unit (square operation units 4-3-1 and 4-4-1, first amplification units 4-5-1 and 4-6-1, and multiplication unit 4). -7-1, 4-8-1, a circuit comprising adders 4-9-1, 4-10-1, and phase rotation units 4-13-1, 4-14-1). Here, the optical phase rotation amount proportional to the square of the amplitude of the optical electric field is returned using the obtained time waveform at the fiber input end.

  At this time, as shown in Equation (2), the phase rotation is the sum of the SPM received from the optical field of its own polarization and the SPM received from the optical field of the polarization orthogonal to itself. Therefore, the SPM compensation unit is a sum of a term proportional to the square of the optical field of its own polarization and a term proportional to the square of the optical field amplitude of the polarization orthogonal to itself.

The optical phase rotation amount requires (1) signal light power, (2) nonlinear coefficient, and (3) nonlinear effective fiber length as proportional coefficients. The unit system is arbitrary, but Leff [km], γ [W ⁻¹ km ⁻¹ ], and signal light power [W] are often used. A typical γ of a single mode fiber is about ˜1 W ⁻¹ km ⁻¹ . Usually, the optical electric field waveform obtained by a digital signal processing circuit for coherent reception basically has no information on optical transmission power, and a standardized value is used in the digital signal processing circuit. Therefore, it is necessary to provide optical transmission power. In the formula (2), it is assumed that the standardization is performed so that the average value of the square of the optical electric field intensity handled by the digital signal processing circuit is 1. Since Ex and _Ey are standardized, _Ps is multiplied to convert the absolute value of the optical power.

  The signal light waveform output from the SPM compensation calculation is then input to the second dispersion compensation calculation section including the chromatic dispersion compensation units 4-1-1 and 4-2-1. In this case, by compensating the wavelength dispersion of the second span counted from the receiving side, the optical electric field time waveform at the input end of the second span counted from the receiving side can be reproduced.

  Next, this optical electric field waveform is converted into a second SPM compensation unit (square calculation units 4-3-2, 4-4-2, first amplification units 4-5-2, 4-6-2, multiplication unit 4). -7-2, 4-8-2, adder units 4-9-2, 4-10-2, and phase rotation units 4-13-2, 4-14-2) The SPM waveform distortion generated in the span is compensated. By repeating this, the optical electric field waveform of the signal light at the input end of the first span counted from the transmission end, that is, the transmission waveform can be obtained. Thereafter, the signal is received through a normal process such as polarization separation using an adaptive FIR filter and phase estimation by the polarization separation circuit 5.

B. Second Embodiment Next, a second embodiment of the present invention will be described.
FIG. 2 is a block diagram showing the configuration of the digital signal processing circuit according to the second embodiment. It should be noted that portions corresponding to those in FIG. In the figure, the dispersion compensation & SPM compensation circuit 10-1 performs dispersion compensation and SPM compensation for X polarization, and the dispersion compensation & SPM compensation circuit 10-2 performs dispersion compensation and SPM compensation for Y polarization. And do.

  The dispersion compensation & SPM compensation circuit 10-1 includes N chromatic dispersion compensation units 10-1-1 to 10-1-N, square calculation units 10-2-1 to 10-2-N, and an amplification unit 10-3-. 1-10-3-N and phase rotation units 10-4-1 to 10-4-N. The dispersion compensation & SPM compensation circuit 10-2 has basically the same configuration.

  The chromatic dispersion compensation unit 10-1-1 receives a digital signal including phase information as input and compensates chromatic dispersion for each polarization. The chromatic dispersion compensation unit 10-1-k (k = 2, 3,..., N) receives the output from the phase rotation unit 10-4- (k-1) of each polarization as input and compensates for chromatic dispersion. . The square computing units 10-2-1 to 10-2-N compute the square of the electric field amplitude of each polarization from the chromatic dispersion compensating units 10-1-1 to 10-1-N in the preceding stage, respectively. Each of the amplification units 10-3-1 to 10-3-N multiplies the output from the square operation units 10-2-1 to 10-2-N in the previous stage by the nonlinear coefficient of the fiber and the nonlinear effective fiber length, respectively. To do. The phase rotation units 10-4-1 to 10-4-N respectively amplify the units 10-3-1 to 10-3-1 with respect to the outputs from the preceding chromatic dispersion compensation units 10-1-1 to 10-1-N. Based on information from 10-3-N, phase rotation is performed.

  The configuration shown in FIG. 2 is characterized in that only the amount of phase rotation due to its own polarization is compensated. Since the SPM received from the orthogonal polarization component is smaller than the SPM received from the polarization of its own, this configuration can sufficiently obtain the effect that the nonlinear effect can be compensated.

C. Third Embodiment Next, a third embodiment of the present invention will be described.
In order to compensate SPM with high accuracy, it is necessary to accurately reproduce the optical electric field waveform in the optical fiber transmission line. As described above, the delay time from the time when the light is input to the 90 ° hybrid from the output end of the optical fiber to the time when the light is input to the ADCs 1-1 to 1-4 is the same phase component of the X polarization component, orthogonal. It differs depending on the four paths of the phase component, the in-phase component of Y polarization, and the quadrature component. For this reason, keeping the skew between them sufficiently small is an important point for accurately reproducing the optical electric field waveform in the optical fiber transmission line.

  In particular, when there is no SPM compensation, even if the skew between the X polarization and the Y polarization remains, it can be compensated by the adaptive FIR filter of the polarization separation circuit 5 at the subsequent stage. Therefore, it is important to reduce the residual skew or to take countermeasures.

  For example, in a 100 Gbps transmission system, since the sampling rate is about 28 GHz, the sampling time interval is about 37 ps. Therefore, in order to obtain a sufficient SPM compensation effect, it is necessary to suppress the time lag between the X polarization and the Y polarization to at least less than about 37 ps.

  FIG. 3 is a block diagram illustrating a configuration in which fine adjustment of the skew adjustment unit is performed by monitoring signal quality in the digital signal processing circuit according to the third embodiment when SPM compensation is performed. In the figure, parts corresponding to those in FIG. The phase estimation circuits 6-1 and 6-2 estimate the phases of the X polarization and the Y polarization. The demodulation circuits 7-1 and 7-2 demodulate according to the estimated phase. The quality monitor 11 monitors the signal quality (for example, Q value) from the output signals from the demodulation circuits 7-1 and 7-2, and supplies the signal quality to the skew control unit 12. The skew control unit 12 controls the skew amount of the skew adjustment units 2-1 to 2-4 so that the signal quality is further improved according to the output signal from the quality monitor 11.

  A constellation diagram plotted on the complex plane can be drawn using the signal after the polarization separation and the phase estimation. This constellation has a characteristic pattern depending on the modulation method.

  FIG. 4 is a conceptual diagram for explaining an example of a constellation diagram in quaternary phase modulation (QPSK). In the example of FIG. 4, the plot can be divided into four groups, which are distributed around four points of π / 4, π3 / 2, 5π / 4, and 7π / 4. The distribution of the point group plotted in this constellation represents the signal quality. As an example in which this spread can be handled quantitatively, there is a Q value estimated from a constellation.

  Also, when a known data pattern or pilot signal is inserted at the transmitting end, the error rate can be measured at the receiving end, so the error rate or the signal quality obtained from the error rate can be used as a quantitative index. it can. The signal quality value is monitored, and the skew is finely adjusted so that the value becomes the highest. For this purpose, the skew control unit 12 gives a minute skew fluctuation, and as a result, the quality monitor 11 monitors whether the signal quality is improved or decreased. Then, the skew is changed in the direction of improving the signal quality.

  FIG. 5 is a conceptual diagram showing an example of the result of evaluating the Q value by changing the skew between the X polarization and the Y polarization. For example, assume that the skew value is set to 15 ps in the initial stage. In this case, when the skew is changed to 10 ps and 20 ps and compared in the two cases, the Q value is improved when set to 10 ps, and the Q value is decreased when set to 20 ps. In this case, the set value is updated to 10 ps.

  In the next step, the skew is changed from 10 ps to 5 ps and 15 ps. In this case, the Q value improves at 5 ps, but the Q value decreases at 15 ps. In this case, the set value is updated to a skew value of 5 ps. By repeating this, an optimal skew can be set. Actually, since there is a measurement error of the Q value and a time variation due to the state of the transmission path, it is desirable to obtain an average value of the Q value by measuring a plurality of times.

D. Fourth Embodiment Next, a fourth embodiment of the present invention will be described.
In the absence of SPM compensation, even if the skew between the X polarization and the Y polarization remains, it can be compensated by the adaptive FIR filter for polarization separation in the subsequent stage. A method for realizing an accurate skew adjustment function in the dispersion compensation & SPM compensation circuit will be described.

  FIG. 6 is a block diagram showing the configuration of the dispersion compensation & SPM compensation circuit of the digital signal processing circuit according to the fourth embodiment. Note that the overall configuration of the digital signal processing circuit is the same as that in FIG. 1, and portions corresponding to those in FIG.

  In the figure, the skew adjustment units 13-1-1-1 to 13-1-N and 13-2-1 to 13-2-N multiply the outputs of the square operation units 4-3-1 to 4-3-N. It is inserted between the sections 4-7-1 to 4-7-N and 4-8-1 to 4-8-N. That is, there is a part that adds the phase rotation contribution from the X polarization component of the SPM compensation unit and the contribution from the SPM phase rotation from the Y polarization component. A skew adjustment function has been added. The skew adjustment method by the skew adjustment units 13-1-1 to 13-1-N and 13-2-1 to 13-2-N is the same as that in the third embodiment described above. The skew value may be adjusted so as to increase.

E. Fifth Embodiment Next, a fifth embodiment of the present invention will be described.
The phase rotation amount by SPM is proportional to the time waveform of the optical electric field of the signal light. As described in the first embodiment, the proportional coefficient is (1) signal light power, (2) nonlinear coefficient, (3) Non-linear effective fiber length. In particular, the nonlinear coefficient depends on the effective cross-sectional area of the transmission line fiber, the fiber material, etc., but the typical value of a fiber used as a transmission line is typically dependent on the fiber type such as dispersion-shifted fiber or single mode fiber. There is no significant shift.

  The nonlinear effective fiber length depends on the loss factor, but there is a typical Leff because the loss factor has a typical value. For example, if the loss coefficient is 0.2 dB / km, Leff is about 21 km. On the other hand, the signal light power is a different parameter depending on the system design, and has a different value each time.

  FIG. 7 is a block diagram showing the configuration of the digital signal processing circuit according to the fifth embodiment. It should be noted that portions corresponding to those in FIG. The digital signal processing circuit shown in FIG. 7 is configured to input signal light power determined by system design as information to the dispersion compensation & SPM compensation circuit 14 and calculate the SPM phase rotation amount using the value.

  Square arithmetic units 4-3-1a to 4--3-Na, 4-3-1b to 4-3-Nb, 4-4-1a to 4--4-Na, 4-4-1b to 4--4-Nb Respectively calculate the square of the electric field amplitude of each polarization from the chromatic dispersion compensation units 4-1-1 to 4-1-N, 4-2-1 to 4-2-N in the preceding stage.

  The amplifying units 4-11-1a to 4-11-Na and 4-12-1a to 4-12-Na are connected to the gain control units 14-1-1a to 14-1-Na and 14-2-1a to 14-, respectively. Based on the gain control signal from 2-Na, its own gain is adjusted, and each polarization is calculated from the square calculation units 4-3-1a to 4--3-Na and 4-4-1a to 4-4-Na. Amplify the output. The amplification units 4-11-1b to 4-11-Nb and 4-12-1b to 4-12-Nb are connected to the gain control units 14-1-1b to 14-1-Nb and 14-2-1b to 14-, respectively. The own gain is adjusted based on the gain control signal from 2-Nb, and is obtained from the square calculation units 4-3-1b to 4--3-Nb and 4-4-1b to 4--4-Nb of each polarization. Amplify the output.

  The phase rotation units 4-13-1a to 4-13-Na and 4-14-1a to 4-14-Na respectively output from the previous stage chromatic dispersion compensation units 4-2-1 to 4-2-N. On the other hand, phase rotation is performed on the basis of information from the amplification units 4-11-1a to 4-11-Na and 4-12-1a to 4-12-Na. The phase rotation units 4-13-1b to 4-13-Nb and 4-14-1b to 4-14-Nb are respectively connected to the preceding phase rotation units 4-13-1a to 4-13-Na and 4-14. -1a to 4-14-Na, the phase rotation is performed based on the information from the amplifiers 4-11-1b to 4-11-Nb and 4-12-1b to 4-12-Nb. Do.

  The gain control units 14-1-1a to 14-1-Na and 14-2-1a to 14-2-Na follow the SPM phase rotation represented by the equation (2), and the value of the input signal light power, and The SPM phase rotation amount obtained by integrating the nonlinear coefficient, the nonlinear effective fiber length, and the like is calculated, and the value is supplied to the amplifying units 4-11-1a to 4-11-Na and 4-12-1a to 4-12-Na. Control the gain of each amplifier. Similarly, the gain control units 14-1-1b to 14-1-Nb and 14-2-1b to 14-2-Nb are used to calculate the input signal light power value, nonlinear coefficient, nonlinear effective fiber length, and the like. The integrated SPM phase rotation amount is calculated, and the value is supplied to the amplifying units 4-11-1b to 4-11-Nb and 4-12-1b to 4-12-Nb, and the gain of each amplifier is controlled. .

  The accuracy with which the signal light power that is actually input to the transmission path can be controlled includes errors due to wavelength variations, absolute optical power measurement accuracy, and the like. Further, other proportional coefficients for determining the SPM rotation amount also differ depending on the conditions of the transmission path. The nonlinear effective fiber length depends on the loss factor, but depending on the optical fiber, the loss varies depending on the year of manufacture, the manufacturer, the state when the fiber is laid, and the like.

  Therefore, the amount of phase rotation by SPM differs depending on the transmission line fiber. The effective cross-sectional area of the nonlinear fiber also has a typical value, but this value also varies depending on the refractive index profile of the cross section of the optical fiber, the manufacturer, the optical fiber core, the material distribution of the cladding, and the like. Further, the optical power changes with time due to an instantaneous change in fiber touch or fiber stress. Therefore, by adaptively adjusting the gain of the SPM compensation unit, it is possible to set a gain that is more suitable for the transmission conditions and to enhance the compensation effect.

F. Sixth Embodiment Next, a sixth embodiment of the present invention will be described.
FIG. 8 is a block diagram showing the configuration of the digital signal processing circuit according to the sixth embodiment. Note that the same reference numerals are given to the portions corresponding to the above-described drawings. In the figure, the gain control units 16-1 to 16-N are first amplifying units 4-5-1 to 4-5 so that the signal quality (for example, Q value) from the quality monitor 11 is increased. -N, 4-6-1 to 4-6-N, multipliers 4-7-1 to 4-7-N, 4-8-1 to 4-8-N are controlled in gain.

  In the sixth embodiment, the quality monitor 11 monitors the signal quality (for example, Q value) from the output signals from the demodulation circuits 7-1 and 7-2, and the gain control unit 16- 1 to 16-N, the first amplifying units 4-5-1 to 4-5-N, 4-6-1 to 4-6-N, multipliers 4-7-1 to 4-7-N, 4 By controlling the gain of -8-1 to 4-8-N, the gain of each SPM compensation unit is adaptively controlled.

  FIG. 9 is a conceptual diagram showing an example of the result of evaluating the Q value by changing the gain in the sixth embodiment. For example, assume that the gain value is set to 0.6 in the initial stage. In this case, when the gain is changed to 0.4 and 0.8 and compared in the two cases, the Q value is improved when set to 0.8, and when set to 0.4, Q value decreases. In this case, the gain setting value is updated to 0.8.

  In the next step, the gain is changed from 0.8 to 0.6 and 1.0. In this case, the Q value improves at 1.0, but the Q value decreases at 0.8. In this case, the gain setting value is updated to 1.0. By repeating this, an optimum gain can be set. Actually, since there is a measurement error of the Q value and a time variation due to the state of the transmission path, it is desirable to obtain an average value of the Q value by measuring a plurality of times.

G. Seventh Embodiment Next, a seventh embodiment of the present invention will be described.
FIG. 10 is a block diagram showing the configuration of the digital signal processing circuit according to the seventh embodiment. In addition, the same code | symbol is attached | subjected to the part with respect to the figure mentioned above, and description is abbreviate | omitted. In the figure, the waveform distortion monitor 20 detects the amount of nonlinear distortion based on the outputs of the phase estimation circuits 6-1 and 6-2 and the outputs of the demodulation circuits 7-1 and 7-2. This is supplied to the compensation circuit 16. The separation compensation & SPM compensation circuit 16 includes gain control units 17-1 to 17-N. The gain control units 17-1 to 17 -N are first amplifying units 4-5-1 to 4-5 -N, 4-6 so that the nonlinear distortion amount from the waveform distortion monitor 20 is minimized. The gains of -1 to 4-6-N, multipliers 4-7-1 to 4-7-N, and 4-8-1 to 4-8-N are controlled.

  As a method for detecting the amount of nonlinear distortion, there is a method for detecting a difference between a target point and an actual measurement point after the phase estimation unit, as shown in FIG. The signal after the polarization separation by the polarization separation circuit 5 and the phase estimation by the phase estimation circuits 6-1 and 6-2 are plotted on a complex plane as shown in FIG. You can draw a figure.

  However, if there is waveform distortion, the constellation distribution deviates from the target point as shown in FIG. This shift is considered to be caused by linear distortion such as chromatic dispersion of the transmission line and polarization mode dispersion, distortion due to nonlinear effects, distortion due to noise, and the like. However, the difference between the non-linear distortion and other distortions is that the non-linear distortion receives a distortion proportional to the amplitude intensity of the waveform. Using this feature, the waveform distortion monitor 20 monitors the amount of nonlinear distortion and feedback controls the gain. The gain control units 17-1 to 17-N calculate the predicted SPM waveform distortion from the time waveforms of the X polarization and the Y polarization.

  FIG. 12 is a block diagram showing the configuration of the gain control units 17-1 to 17-N of the digital signal processing circuit according to the seventh embodiment. In the figure, the gain control units 17-1 to 17-N include a distortion estimation unit 18-1, a division unit 18-2, an averaging unit 18-3, an addition unit 18-4, and a delay unit 18-5. The distortion estimation unit 20-1 predicts a theoretically predicted distortion. For example, the SPM distortion generated in the X-polarized signal light electric field can be approximately expressed by the following equation (3).

  Accordingly, the waveform distortion can be estimated using the time waveforms of the optical fields of the X polarization and the Y polarization output from the kth dispersion compensation section. The time waveform of the optical electric field output from the k-th dispersion compensation section is not received as it is, and there are (N−k) chromatic dispersion compensation units in the subsequent stage.

  For example, as shown in FIG. 13, the phase distortion estimation unit 4-21-k estimates the waveform distortion from the output from the chromatic dispersion compensation unit 4-1-k, and the waveform distortion calculation unit 4-23-k performs chromatic dispersion. The waveform distortion is calculated from the output from the compensation unit 4-1-k and the estimated waveform distortion, and the residual wavelength dispersion compensation unit 4-25-k compensates the residual wavelength dispersion. As described above, when the distortion amount is estimated using the optical electric field waveform of the kth dispersion compensation section, the remaining amount received by the (N−k) dispersion compensation sections with respect to the obtained waveform distortion amount. The residual dispersion compensation operation corresponding to the dispersion compensation is applied. In FIG. 13, only the SPM phase rotation received from the waveform of the X polarization is estimated as the waveform distortion by the phase distortion estimation unit 4-21-k, but the waveform distortion received from the Y polarization is also shown in FIG. It can be similarly estimated with the configuration as shown in FIG.

  Next, a correlation between the waveform distortion calculated by the distortion estimation unit 18-1 and the actually measured waveform distortion amount ΔE is detected. First, the division unit 18-2 divides the measured waveform distortion amount ΔE by the estimated estimated distortion (Y / X) to calculate a proportional coefficient between them. Since a noise component is superimposed on the actually measured signal, a more reliable proportionality coefficient can be calculated by taking the average in the averaging unit 18-3. The proportional coefficient is added to the current set value to obtain a gain at the next time. In this configuration, since the correlation between the actual distortion amount and the estimated distortion amount is fed back via the delay unit 18-5 and the addition unit 18-4, it is updated in the direction of reducing the distortion. Approaching a suitable gain.

  The proportionality coefficient calculated by this method includes an error due to the influence of noise or the like. Therefore, when updating the gain, taking this into consideration, the gain increment actually calculated is multiplied by a value smaller than 1 called the step size, and the update speed is slowed down, giving a sudden large change. It is preferable to gradually update the reception Q value so as not to deteriorate the reception Q value.

H. Eighth Embodiment Next, an eighth embodiment of the present invention will be described.
FIG. 14 is a block diagram showing the configuration of the digital signal processing circuit according to the eighth embodiment. In addition, the same code | symbol is attached | subjected to the part with respect to the figure mentioned above, and description is abbreviate | omitted. FIG. 14 shows only the kth chromatic dispersion compensation unit and the SPM compensation unit, but the other sections have the same configuration. The chromatic dispersion compensator 4-1 -k receives a digital signal including phase information as input and compensates chromatic dispersion for each polarization. The square calculation unit 30-1 calculates the square of the electric field amplitude of each polarization from the chromatic dispersion compensation units 30-1 to 10-1-N in the previous stage.

  The square calculation units 30-2 to 30-4 calculate the squares of the outputs from the FIR filters 32-1 to 32-3, respectively. The amplification units 31-1 to 30-4 amplify the outputs from the square calculation units 30-1 to 30-4, respectively. The adders 33-1 to 33-5 sequentially accumulate outputs from the square calculators 30-1 to 30-4. The phase rotation unit 4-13-k performs phase rotation on the output from the previous stage chromatic dispersion compensation unit 4-1-k based on the information from the final stage addition unit 33-1.

  In the eighth embodiment, part of the chromatic dispersion compensation is replaced with FIR filters 25-1, 25-2, and 25-3, so that the number of divisions of compensation calculation is equivalently increased, and the division unit of chromatic dispersion compensation is changed. By making it finer, SPM waveform distortion can be compensated more accurately.

I. Ninth Embodiment Next, a ninth embodiment of the present invention will be described.
FIG. 15 is a block diagram showing a configuration of a digital signal processing circuit according to the ninth embodiment. In addition, the same code | symbol is attached | subjected to the part with respect to the figure mentioned above, and description is abbreviate | omitted. FIG. 15 shows only the kth chromatic dispersion compensation unit and the SPM compensation unit, but the other sections have the same configuration.

  The chromatic dispersion compensators 4-1 -k, 4-2 -k (k = 2, 3,..., N) receive the outputs from the phase rotation units 4-14-(k−1) of the respective polarizations as inputs. The chromatic dispersion is compensated for each polarization. The square calculation unit 36-1-ka calculates the square of the electric field amplitude of the X polarization from the chromatic dispersion compensation unit 4-1-k in the previous stage. The FIR filter 35-1-k performs filtering on the X polarized wave in which the chromatic dispersion from the chromatic dispersion compensating unit 4-1-k is compensated. The square calculation unit 36-1-kb calculates the square of the output from the FIR filter 35-1-k.

  Each of the amplifying units 37-1-ka and 37-1-kb converts the nonlinear coefficient of the X-polarized fiber and the nonlinear effective fiber length into the X-polarized square computing units 36-1-ka, 36-1- Multiply the output from kb. Each of the multipliers 38-1-ka and 38-1-kb multiplies the nonlinear coefficient of the other polarization fiber and the nonlinear effective fiber length. The adder 39-1-kb adds the output of the amplifier 37-1-kb, the output of the multiplier 38-2-ka corresponding to the other polarization, and the output of the multiplier 438-2-kb. . The adder 39-1-ka adds the addition output from the adder 39-1-kb and the output of the amplifier 37-1-ka, and supplies the result to the phase rotation unit 4-13-k. The phase rotation unit 4-13-k performs phase rotation on the output from the previous stage chromatic dispersion compensation unit 4-1-k based on the information from the addition unit 39-1-ka.

  On the other hand, the square calculation unit 36-2-ka calculates the square of the electric field amplitude of the Y polarization from the chromatic dispersion compensation unit 4-2-k in the previous stage. The FIR filter 35-2-k performs filtering on the Y polarized wave in which the chromatic dispersion from the chromatic dispersion compensating unit 4-2k is compensated. The square calculator 36-2-kb calculates the square of the output from the FIR filter 35-2-k. Each of the amplifying units 37-2-ka and 37-2-kb converts the nonlinear coefficient and the nonlinear effective fiber length of the Y-polarized fiber into the Y-polarized square calculation units 36-2-ka and 36-2-, respectively. Multiply the output from kb.

  Each of the multipliers 38-2-ka and 38-2-kb multiplies the nonlinear coefficient of the other polarization fiber by the nonlinear execution fiber length. The adder 39-2-kb adds the output of the amplifier 37-2-kb, the output of the multiplier 38-1-ka corresponding to the other polarization, and the output of the multiplier 38-1-kb. . The adding unit 39-2-ka adds the addition output from the adding unit 39-2-kb and the output of the amplifying unit 37-2-ka, and supplies the result to the phase rotating unit 4-14-k. Each of the phase rotators 4-14-k performs phase rotation on the output from the preceding chromatic dispersion compensator 4-2-k based on the information from the adder 39-2-ka.

  In the ninth embodiment, as shown in FIG. 15, FIR filters 35-1-k and 35-2-k are inserted in the SPM compensation unit, and the time waveform of phase rotation given for SPM compensation is provided. Is a method of optimizing Originally, the waveform change due to chromatic dispersion and the waveform change due to SPM proceed simultaneously during propagation through the optical fiber transmission line. On the other hand, these compensation operations are approximately compensated by a configuration in which chromatic dispersion compensation and SPM compensation are alternately repeated. Therefore, the SPM waveform distortion can be compensated more accurately by making the division unit of chromatic dispersion compensation finer. However, if the division unit of dispersion compensation is made finer, the amount of calculation increases, so it is effective with as little calculation amount as possible. It is important to realize the SPM compensation effect.

  In addition, in order to amplify the optical power in the optical amplifier and input it to the optical fiber, if it is sandwiched between two optical amplifiers, the nonlinear effect at the input end of the optical fiber transmission section (span) is generally large, It is effective to perform SPM compensation at the input end. Therefore, it is effective to set the dispersion compensation division unit for each span.

  However, there is a Raman amplification technique in which pumping light is input from the output end or the input end of the optical fiber transmission line and is amplified using the Raman effect of the optical fiber transmission line. In this case, the signal light power is high not only at the input end of each span but also at the output end. Therefore, when using Raman amplification, it is considered necessary to make the division unit for dispersion compensation finer. However, in this case as well, there is a problem in that the amount of calculation increases if the dispersion compensation division unit is made fine.

  In the configuration example of FIG. 15, the output waveform from the k-th stage chromatic dispersion compensator 4-1-k, 4-2k is inserted into the FIR filters 35-1-k, 35-2-k, and transmitted. A waveform in which the SPM waveform distortion is most strongly generated in the road is reproduced, and a phase rotation proportional to the square of the amplitude of the waveform is given in the reverse direction. For example, the waveform input from the k-th stage is generated by the square calculation units 36-1-ka and 36-2-ka as it is, generating a square waveform of the amplitude as it is and using it as a part of the phase rotation amount. Used to return the phase of the polarization signal waveform.

  The waveform input from the k-th stage is input to the FIR filters 35-1-k and 35-2-k. The FIR filters 35-1-k and 35-2-k have FIR filter tap coefficients set so as to compensate for the chromatic dispersion amount of a part of the span, for example, half the chromatic dispersion amount. As a result, a time waveform traced back further to the transmission side can be generated for the next half span. This time waveform is squared by the square calculation units 36-1-kb and 36-2-kb to generate a time waveform of SPM phase rotation, and this is also used as a part of the k-th phase rotation amount. Used to return the phase of the wave signal waveform.

  In the phase rotation units 4-13-k and 4-14-k, when giving phase rotation, the weights are added after applying appropriate weights. Thereby, as a time waveform of the square of the optical electric field used for SPM compensation, a plurality of time waveforms are prepared at a division level finer than the division unit of the chromatic dispersion compensation units 4-1-k and 4-2-k. Using this, SPM phase compensation can be realized. Therefore, the chromatic dispersion compensation unit is approximately divided without increasing the number of the chromatic dispersion compensation units 4-1-k, 4-2-k and the phase rotation units 4-13-k and 4-14-k. It is possible to achieve an effect that is made finer.

J. et al. Tenth Embodiment Next, a tenth embodiment of the present invention will be described.
In the configuration described above, by setting the gain of each amplification unit, a weight suitable for the phase rotation time waveform output from each FIR filter is applied, and the phase rotation amount for total SPM compensation is generated. Therefore, it is necessary to control these gains. In the tenth embodiment, the gain is feedback controlled so that the signal quality is monitored and maximized, or the nonlinear distortion amount is detected and the distortion amount is minimized.

  In the method of controlling to maximize the signal quality, the gain control unit receives the demodulated signal quality value detected by the quality monitor, and controls the gain so that the signal quality becomes high. As a method for quantitatively expressing the signal quality, there is a method using a constellation.

  FIG. 16 is a block diagram showing the configuration of the digital signal processing circuit according to the tenth embodiment. In addition, the same code | symbol is attached to the part with respect to the figure mentioned above, and description is abbreviate | omitted. FIG. 16 shows only the kth chromatic dispersion compensation unit and the SPM compensation unit, but the other sections have the same configuration. The gain control unit 40 controls the gain so that the signal quality monitored by a quality monitor (not shown) is maximized, or the nonlinear distortion amount is detected and the distortion amount is minimized.

  As a method for quantitatively expressing signal quality, when constellation is used, as shown in FIG. 4 described above, the spread of the constellation distribution representing the signal quality is quantitatively expressed from the constellation diagram. For example, an optimal gain may be set based on a possible Q value estimated from the constellation. Actually, since there is a measurement error of the Q value and a time variation due to the state of the transmission path, it is desirable to obtain an average value of the Q value by measuring a plurality of times.

  It is also possible to detect the nonlinear distortion amount and control the gain so that the distortion amount is minimized. As a method for detecting the amount of nonlinear distortion, in the subsequent stage of the phase estimation unit, as shown in FIG. 11 described above, the difference between the target point and the actual measurement point in the constellation distribution is detected to detect the amount of nonlinear distortion. The gain is controlled so that the amount of distortion is minimized.

  The gain controller 40 calculates the predicted SPM waveform distortion from the time waveforms of the X polarization and the Y polarization. Note that the SPM waveform distortion prediction by the gain control unit 40 is calculated by the configuration shown in FIG. Accordingly, the waveform distortion can be estimated using the time waveforms of the optical fields of the X polarization and the Y polarization output from the kth dispersion compensation section.

  As described above, the proportionality coefficient calculated by this method includes an error due to the influence of noise or the like. Therefore, when updating the gain, taking this into consideration, the gain increment actually calculated is multiplied by a value smaller than 1 called the step size, and the update speed is slowed down, giving a sudden large change. It is preferable to gradually update the reception Q value so as not to deteriorate the reception Q value.

K. Application Example Next, effects when the present invention is applied to reception of orthogonal frequency division multiplexing (OFDM) signal light will be described.
FIG. 17 is a block diagram showing a configuration of a receiver when the present invention is used for reception of orthogonal frequency division multiplexing (OFDM) signal light. In orthogonal frequency division multiplexing (OFDM), a transmitter generates signal light obtained by orthogonal frequency multiplexing two subcarriers in which each subcarrier is quaternary phase modulated (QPSK) with a modulation frequency of B [GHz]. . The frequency interval between subcarriers is set to B [GHz]. This is coherently received after transmission through a single mode fiber.

  In the figure, the receivers are the ADCs 50-1-1 and 50-1-2, the resample circuit 51-1, the dispersion compensation & SPM compensation circuit 52-1, the carrier separation circuit 53-1, and the X polarization. A frequency control circuit 54-1, an FIR filter 55-1, a carrier phase recovery circuit 56-1, a QPSK decoding circuit 57-1, an FIR filter 58-1, and a QPSK decoding circuit 59-1, ADC 50-2-1, 50-2-2, resampling circuit 51-2, dispersion compensation & SPM compensation circuit 52-2, carrier separation circuit 53-2, automatic frequency control circuit 54-2, FIR filter (12 tap, CMA) 55-2, a carrier phase recovery circuit 56-2, a QPSK decoding circuit 57-2, an FIR filter (12 tap, LMA) 58-2, and a QPSK decoding circuit 59-2.The function of each unit of the transmitter is the same as that of an existing orthogonal frequency division multiplexing (OFDM) transmitter, and thus the description thereof is omitted.

  Here, in the dispersion compensation & SPM compensation circuits 52-1 and 52-2, the dispersion compensation units 60-1, 60-2,..., 60 -N are divided into dispersion amounts for each span. SPM compensation units 70-1, 70-2,..., 70-N described in any of the first to tenth embodiments are inserted.

  FIG. 18 is a conceptual diagram showing an effect when used for reception of orthogonal frequency division multiplexing (OFDM) signal light. FIG. 18 shows the received signal quality Q value when the signal light power input to the transmission line fiber is changed. In the case where there is no compensation, it can be seen that when the transmission power is increased, nonlinear waveform deterioration occurs during fiber propagation, and the signal quality deteriorates. It can be seen that the reception Q value at each transmission power is improved by applying the compensation.

L. Next, a modification according to the present invention will be described.
In the above-described embodiment, the dispersion compensation filter is multistaged by a method in which the chromatic dispersion compensation filter and the SPM phase rotation compensation calculation are only repeated. Therefore, if the accuracy of dispersion compensation required for each dispersion compensation filter is low, the error is multistaged and an error occurs in the total compensation amount. Accordingly, the accuracy with respect to dispersion compensation is increased, and the number of taps of each dispersion compensation filter is increased.

  For example, in order to implement chromatic dispersion and SPM compensation for transmission of several thousand kilometers using a configuration that simply repeats the dispersion compensation filter and SPM phase rotation compensation calculation, individual dispersion compensation for compensating chromatic dispersion of 80 km is used. When the number of filter taps was set to 128, the signal quality Q value was deteriorated. When the number of taps was set to 256 or more, there was a phenomenon in which the deterioration of the Q value was eliminated. When such a filter having a large number of taps is multi-staged, an increase in circuit scale is inevitable. Therefore, the number of taps of the filter of the chromatic dispersion compensation unit is reduced, and the circuit scale by multiple stages is reduced by using a small number of tap filters.

  In order to reduce the amount of calculation and the circuit scale, the configuration shown in FIG. 19 is proposed. Here, the number of taps of the dispersion compensation filter to be repeatedly used is reduced, and the minority tap filter is used to prevent an increase in circuit scale due to multiple stages. On the other hand, this is a method in which errors of dispersion compensation amounts generated by reducing the number of taps of each dispersion compensation filter are collectively compensated by a multi-tap filter. FIG. 20 is a configuration diagram when performing SPM compensation calculation between polarized waves.

  FIG. 19 is a block diagram showing a configuration of a digital signal processing circuit according to this modification. In addition, the same code | symbol is attached | subjected to the part with respect to the figure mentioned above, and description is abbreviate | omitted. In this modification, in order to reduce the calculation amount and the circuit scale, as shown in FIG. 19, minority tap filters 80-1-1 to 80-1 -N with a reduced number of taps are provided as chromatic dispersion compensation units. By repeatedly using the minority tap filters 80-1-1-1 to 80-1-N, an increase in circuit scale due to multiple stages is prevented.

  On the other hand, this is a method in which errors of dispersion compensation amounts generated by reducing the number of taps of each dispersion compensation filter are collectively compensated by a multi-tap filter. A configuration using a time-domain filter is assumed as the minority tap filter.

  FIG. 20 is a block diagram showing a configuration of a digital signal processing circuit in the present modification when performing SPM compensation calculation between polarized waves. In the figure, a small number of tap filters 80-1-1-1 to 80-1-N, 80-2-1 to 80-2-N with a reduced number of taps are provided as chromatic dispersion compensation units, and a large number of taps with an increased number of taps are provided. Tap filters 90-1 and 90-2 are provided in the final stage. The majority tap filters 90-1 and 90-2 are dispersion compensation amounts generated by reducing the number of taps of the minority tap filters 80-1-1 to 80-1 -N and 80-2-1 to 80-2 -N. Compensate for errors together.

As shown in FIG. 21, a configuration using a time domain filter is assumed as the minority tap filter. In the example of FIG. 21 and using the FIR filter 100-1 to 100-N of the N taps, respective weight _C 1 on the data time shift N symbols, ..., the coefficients _{C N,} the multiplication unit 101- Multiply by 1-101-N and add by adders 102-1 to 102- (N-1). When this discrete time response function is represented by a number M samples larger than N, R (n) _span-k = {0,..., 0, 0, C _N ,..., C ₂ , C ₁ , 0, 0, 0 ,..., 0}.

  The response function when the number of minority tap filters is multi-staged is the product of the frequency response function H (n) _span-k obtained by convolution of each minority tap filter or Fourier transform H (n) _span-1 × H (n) _span. −2 ×... × H (n) _span−k. R (n) _total = {d1, d2, d3,..., DM} is an ideal response function for total dispersion compensation in a certain section where dispersion compensation and SPM compensation are performed, and the Fourier transform thereof is H (n ) _Total, tap coefficients of the multi-tap filters 90-1 and 90-2 are H (n) _total-H (n) _span-1 × H (n) _span-2 ×... × H (n) _span− A tap coefficient corresponding to the frequency response function represented by k is set. As a result, the compensation errors generated by the minority tap filters 80-1-1 to 80-1 -N and 80-2-1 to 80-2 -N can be absorbed by the majority tap filters 90-1 and 90-2.

As a configuration method of the multi-tap filter, there are a method using the time domain equalization filter shown in FIG. 21 and a method using a frequency domain equalization filter using FFT (Fast Fourier Transform) as shown in FIG. . In FIG. 22, an S / P converter 120 that converts a serial signal into a parallel signal, an FFT 121 that performs a fast Fourier transform, an IFFT (Inverse FFT) 122 that performs an inverse fast Fourier transform, and an S / P that converts a parallel signal into a serial signal. in addition to the conversion unit 123, the output from FFT121, as described above, respective weight _C 1, ..., the coefficients _{C N,} may be multiplied by the multiplying unit 130-1 to 130-N.

  The use of both can be reduced by using a time domain equalization filter when the number of taps is small and using a frequency domain equalization filter when the number of taps is large.

1-1 to 1-4 ADC
2-1 to 2-4, 13-1-1 to 13-1-N, 13-2-1 to 13-2-N Skew adjustment unit 3-1, 3-2 complex synthesis circuit 4, 10-1, 10-2, 16 Dispersion Compensation & SPM Compensation Circuit 4-1-1 to 4-1-N, 4-2-1 to 4-2-N, 10-1-1 to 10-1-N, 4-1 k, 4-2-k chromatic dispersion compensation unit 4-3-1 to 4-3-N, 4-4-1 to 4-4-N, 10-2-1 to 10-2-N, 30-1 -30-4, 36-1-ka, 36-1-kb, 36-2-ka, 36-2-kb Square operation unit 4-5-1 to 4-5-N, 4-6-1 to 4 -6-N first amplifying unit 4-7-1 to 4-7-N, 4-8-1 to 4-8-N first multiplying unit 4-9-1 to 4-9-N, 4 -10-1 to 4-10-N, 33-1 to 33-4, 39-1-ka, 39-1- b, 39-2-ka, 39-2-kb Adder 4-13-1 to 4-13-N, 4-14-1 to 4-14-N, 10-4-1 to 10-4-N , 4-1-k, 4-13-k, 4-14-k Phase rotation unit 5 Polarization separation circuit 11 Quality monitor 12 Skew control unit 14-1-1a to 14-1-Na, 14-1-1b -14-1-Nb, 14-2-1a to 14-2-Na, 14-2-1b to 14-2-Nb, 16-1 to 16-N, 17-1 to 17-N Gain control unit 18 -1 Distortion estimation unit 18-2 Division unit 18-3 Averaging unit 18-4 Addition unit 18-5 Delay unit 20 Waveform distortion monitor 32-1 to 32-3, 35-1-k, 35-2-k FIR Filter 37-1-ka, 37-1-kb, 37-1-ka, 37-1-kb Amplifying unit 38-1-ka, 38-1-kb 38-2-ka, 38-2-kb Multiplying unit 40 Gain control unit 60-1 to 60-N Dispersion compensation unit 70-1 to 70-N SPM compensation unit 80-1-1 to 80-1-N Few taps Filter 80-2-1 to 80-2-N Minority tap filter 90 Multiple tap filter 90-1 to 90-2 Multiple tap filter 100-1 to 100-N FIR filter 101-1 to 101-N Multiplier 102-1 102- (N-1) Adder 120 S / P converter 121 FFT
122 IFFT
123 S / P converter 130-1 to 130-N multiplier

Claims (11)

A digital signal processing circuit in an optical fiber transmission system,
n (n is a natural number) block compensation arithmetic circuit,
The first compensation arithmetic circuit is
A first chromatic dispersion compensation unit that receives a digital received signal including phase information as input and compensates for chromatic dispersion;
A first square calculation unit that calculates the square of the electric field amplitude from the output of the first chromatic dispersion compensation unit;
A first amplifying unit that multiplies the output from the first squaring unit by a nonlinear coefficient of the optical fiber and a nonlinear effective fiber length;
A first phase rotation unit that performs phase rotation on the output signal from the first chromatic dispersion compensation unit based on output information from the first amplification unit;
The k-th (k = 2 to n natural number) compensation arithmetic circuit is
A kth chromatic dispersion compensator that compensates for chromatic dispersion with respect to an output from the (k−1) th phase rotation unit;
A kth square computing unit that computes the square of the electric field amplitude from the output of the kth chromatic dispersion compensation unit;
A k-th amplification unit that multiplies the output from the k-th square calculation unit by the nonlinear coefficient of the optical fiber and the nonlinear effective fiber length;
A k-th phase rotation unit that performs phase rotation on the output signal from the k-th chromatic dispersion compensation unit based on output information from the k-th amplification unit,
A digital signal processing circuit. A digital signal processing circuit in an optical fiber transmission system,
n (n is a natural number) block compensation arithmetic circuit,
The first compensation arithmetic circuit is
A first chromatic dispersion compensator that compensates chromatic dispersion for each polarization with respect to a digital received signal that is polarized and includes phase information;
A first square calculator that calculates the square of the electric field amplitude of each polarization from the output of the first chromatic dispersion compensator;
A first amplifying unit that multiplies the output from the first square computing unit of each polarization by the nonlinear coefficient and the nonlinear effective fiber length of each polarization fiber;
A first multiplier for multiplying the nonlinear coefficient of the other polarization fiber by the nonlinear effective fiber length;
A first two-input adder that adds the output of the first amplifier and the output of the first multiplier of the other polarization;
A first phase rotation unit that performs phase rotation on the output signal from the first chromatic dispersion compensation unit based on output information from the first two-input addition unit;
The k-th compensation operation circuit (k is a natural number from 2 to n) is
A kth chromatic dispersion compensator for compensating chromatic dispersion for each polarization with respect to an output from the (k−1) th phase rotation unit of each polarization;
A kth square calculation unit that calculates the square of the electric field amplitude of each polarization from the output of the kth chromatic dispersion compensation unit;
A k-th amplification unit that multiplies the output from the k-th square operation unit of each polarization by the nonlinear coefficient and the nonlinear effective fiber length of each polarization fiber;
A kth multiplier for multiplying the nonlinear coefficient of the other polarization fiber by the nonlinear effective fiber length;
A kth two-input adder for adding the output of the kth amplifier and the output of the kth multiplier of the other polarization;
A k-th phase rotation unit that performs phase rotation on the output signal from the k-th chromatic dispersion compensation unit based on output information from the k-th two-input addition unit;
A digital signal processing circuit. For each n block compensation arithmetic circuit,
A (an integer greater than or equal to 1) FIR filters connected in cascade;
A number of square arithmetic units that square the output of each of the a number of FIR filters;
An all-input adder that adds all the outputs of the first square calculator or the k-th square calculator and the outputs from the a square operators;
The a FIR filters are:
A first FIR filter that performs a convolution operation on the output of the kth chromatic dispersion compensator;
It comprises (a-1) second to a-th FIR filters that perform a convolution operation on the output of the preceding stage FIR filter connected in cascade,
The k-th phase rotation unit performs phase rotation on the output signal from the k-th chromatic dispersion compensation unit based on output information from the all-input addition unit;
The digital signal processing circuit according to claim 1. For each n block compensation arithmetic circuit,
An L-th FIR filter that performs a convolution operation on each polarization of the output of the L-th (L = 1 to n) chromatic dispersion compensation unit;
A square calculator that calculates the square of the electric field amplitude of each polarization from the output of the L-th FIR filter;
An amplification unit that multiplies the output from the square calculation unit of each polarization by the nonlinear coefficient and nonlinear effective fiber length of the fiber of each polarization,
A multiplier for multiplying the nonlinear coefficient of the other polarization fiber by the nonlinear effective fiber length;
A three-input adder for adding the output of the amplifier, the output of the multiplier of the other polarization, and the output of the Lth multiplier of the other polarization;
The L-th 2-input adder is
Adding the output of the L-th amplifier and the output of the 3-input adder;
The L-th phase rotation unit is
Based on the output information from the L-th two-input adder, the polarization-separated digital received signal including phase information or the output signal from the (L-1) -th chromatic dispersion compensator , Phase rotation,
The digital signal processing circuit according to claim 2. The first to kth chromatic dispersion compensators are as follows:
Featuring a minority tap filter with fewer taps,
A multi-tap filter with an increased number of taps, which performs predetermined signal processing on the output signal of the M-th (= n) -th phase rotation unit and compensates for errors in chromatic dispersion compensation by the small-tap filter;
The digital signal processing circuit according to claim 1. The first to kth chromatic dispersion compensators are as follows:
Featuring a minority tap filter with fewer taps,
A multi-tap filter with an increased number of taps, which performs predetermined signal processing on the output signal of the M-th (= n) -th phase rotation unit and compensates for errors in chromatic dispersion compensation by the small-tap filter;
The digital signal processing circuit according to claim 2.   7. The gain control unit according to claim 1, further comprising: a gain control unit configured to control a gain of an input signal to the first to k-th phase rotation units based on a predetermined signal light power. Digital signal processing circuit. When the dispersion compensation amount of the k-th chromatic dispersion compensator is set to 1, 2,..., K in order from the reception end of the transmission section sandwiched between two optical amplifiers, the chromatic dispersion amount in the k-th section is obtained. The digital signal processing circuit according to claim 1, wherein the digital signal processing circuit is set.
An optical receiver that includes the digital signal processing circuit according to any one of claims 1 to 6 and that receives an optical signal having X polarization and Y polarization,
A skew adjustment unit that adjusts the skews of the in-phase component and the quadrature component of the X-polarized wave and the Y-polarized wave;
A quality monitoring unit for monitoring the signal quality after demodulation of the optical signal;
An optical receiver comprising: a control unit that controls a skew amount of the skew adjustment unit based on signal quality information output from the quality monitor unit. The quality monitor unit
A waveform distortion estimator that estimates changes in the optical electric field waveform due to self-phase modulation from the received signal;
A waveform distortion detector for detecting a waveform distortion amount of the optical electric field waveform from the demodulated signal of the optical signal,
The controller is
Based on the difference between the change in the optical electric field waveform estimated by the waveform distortion estimation unit and the waveform distortion amount of the optical electric field waveform detected by the waveform distortion detection unit, the skew amount of the skew adjustment unit is controlled.
The optical receiver according to claim 9. The quality monitor unit
A waveform distortion estimator that estimates changes in the optical electric field waveform due to self-phase modulation from the received signal;
A waveform distortion detector for detecting a waveform distortion amount of the optical electric field waveform from the demodulated signal of the optical signal,
The controller is
Based on the difference between the change in the optical electric field waveform estimated by the waveform distortion estimation unit and the waveform distortion amount of the optical electric field waveform detected by the waveform distortion detection unit, the in-phase components of the X polarization and the Y polarization , To control the gain that determines the amount of phase rotation of the quadrature component,
The optical receiver according to claim 9.

Images