Signal processing method, device and coherent receiver

Abstract

The present application provides a signal processing method, apparatus, and coherent receiver. The signal processing method includes obtaining P real-number signals, subjecting the P real-number signals to at least number-theoretic transform (NTT) processing to obtain a first real-number signal in the P transform domain, performing at least clock recovery on the first real-number signal in the P transform domain to obtain a second real-number signal in the P transform domain, subjecting the second real-number signal in the P transform domain to at least polarization compensation and inverse number-theoretic transform (INTT) processing to obtain m complex-number signals X and Y in the time domain, and performing phase recovery and decoding on the m complex-number signals X and Y in the time domain to obtain bit signals. Using NTT and INTT processing in the signal processing process can improve signal processing performance and reduce complexity.

Description

Technical Field

[0001] The present application relates to the field of optical communications, and in particular to a signal processing method, device and coherent receiver. Background Art

[0002] The performance of high-capacity optical communication systems may be degraded by fiber attenuation, fiber dispersion, polarization mode dispersion, laser phase noise, and fiber nonlinearity. Currently available polarization-multiplexed coherent receivers can effectively compensate for these degradation factors in the electrical domain during electrical signal processing.

[0003] To achieve effective compensation in the electrical domain, existing polarization-multiplexed coherent receivers process electrical signals by convolving the time-domain signal with the inverse function of the channel response based on Fourier transforms. This requires the introduction of fast Fourier transform (FFT) and inverse fast Fourier transform (IFFT) modules into the polarization-multiplexed coherent receiver architecture. However, the high complexity and low precision of Fourier transforms lead to high complexity in the electrical signal processing and low accuracy in the recovered signals. Therefore, improving the performance of polarization-multiplexed coherent receivers in processing electrical signals has become an urgent issue. Summary of the Invention

[0004] The present application provides a signal processing method, device and coherent receiver to improve the performance of signal processing.

[0005] In a first aspect, a signal processing method is provided. The signal processing method can be performed by a coherent receiver, or can also be performed by a chip or circuit provided in the coherent receiver, which is not limited in this application.

[0006] The signal processing method includes:

[0007] First, obtaining P real-number signals, where the P real-number signals include P real-number signals after analog-to-digital conversion or other P real-number signals that require digital signal processing, which is not limited in this application;

[0008] Then, performing at least a number theoretic transform (NTT) process on the P-way real-number signal to obtain a first real-number signal in a P-way transform domain;

[0009] Then, at least clock recovery is performed on the first real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain;

[0010] Secondly, the second real number signal in the P-way transform domain is subjected to at least polarization compensation and inverse number theoretic transform (INTT) processing to obtain m-way complex number signals X in the time domain and m-way complex number signals Y in the time domain, where m and P are positive integers.

[0011] Furthermore, if INTT and merging processing are performed first and then polarization compensation processing is performed, the signal processing method provided in the present application can perform phase recovery and decoding on the m-channel complex signal X in the time domain and the m-channel complex signal Y in the time domain after polarization compensation to obtain a bit signal;

[0012] It should be understood that if the INTT and merging processes are performed first and then the polarization compensation process is performed, the polarization compensation process is performed based on the time domain signal after the INTT process. Specifically, performing polarization compensation in the time domain can enhance the ability to resist loop delay.

[0013] Alternatively, if polarization compensation processing is performed first and then INTT and merging processing is performed, the signal processing method provided in the present application can perform phase recovery and decoding on the m-path complex signal X in the time domain in the first polarization direction and the m-path complex signal Y in the time domain in the second polarization direction after the above-mentioned merging to obtain a bit signal, thereby realizing signal recovery.

[0014] It should be understood that polarization compensation is performed first and then INTT and merging is performed. In other words, polarization compensation is performed based on signals in the transform domain. Specifically, performing polarization equalization damage compensation in the transform domain and replacing convolution with multiplication can reduce power consumption.

[0015] The signal processing method provided in the present application obtains signals in two polarization directions through analog-to-digital conversion. The signals in the two polarization directions are input into a receiving digital signal processing Rx DSP device. During the processing of the Rx DSP, number theoretic transform NTT and inverse number theoretic transform INTT are used instead of fast Fourier transform FFT and inverse fast Fourier transform IFFT, thereby avoiding the problems of high complexity and low accuracy brought by FFT and IFFT processing, thereby improving the performance of signal processing.

[0016] In conjunction with the first aspect, in certain implementations of the first aspect, the P-way real-number signal is subjected to at least number theoretic transform (NTT) processing to obtain the first real-number signal in the P-way transform domain, including: subjecting the P-way time-domain signal to NTT processing to obtain the first real-number signal in the P-way transform domain; or

[0017] The P-channel real number signal is processed by digital back propagation DBP to obtain the tenth real number signal in the P-channel time domain; the tenth real number signal in the P-channel time domain is processed by NTT to obtain the first real number signal in the P-channel transform domain.

[0018] Among them, if P-channel real-number signals are processed by DBP, the DBP module is used to compensate for the nonlinear effect and increase the signal transmission distance.

[0019] In the signal processing method provided herein, the first real-number signal in the P-channel transform domain can be obtained by inputting the P-channel real-number signal into an NTT module and performing NTT processing. Alternatively, the first real-number signal in the P-channel transform domain can be obtained by first inputting the P-channel real-number signal into a DBP module, performing DBP processing, and then performing NTT processing. These different methods for obtaining the first real-number signal in the P-channel transform domain provide greater flexibility in the structural design of the Rx DSP device.

[0020] In combination with the first aspect, in certain implementations of the first aspect, when the P-path real-number signal has not undergone DBP processing, the above-mentioned obtaining of the second real-number signal in the P-path transform domain based on the first real-number signal in the P-path transform domain needs to undergo dispersion compensation, that is, the first real-number signal in the P-path transform domain is at least subjected to clock recovery, and obtaining the second real-number signal in the P-path transform domain includes: performing dispersion compensation on the first real-number signal in the P-path transform domain to obtain a third real-number signal in the P-path transform domain; performing clock recovery on the third real-number signal in the P-path transform domain to obtain the second real-number signal in the P-path transform domain; or,

[0021] When the P-path real number signal is processed by DBP, since the DBP module has the function of dispersion compensation, the above-mentioned method of obtaining the second real number signal in the P-path transform domain based on the first real number signal in the P-path transform domain does not require additional dispersion compensation, that is, the first real number signal in the P-path transform domain is at least clock recovered, and obtaining the second real number signal in the P-path transform domain includes: performing clock recovery on the first real number signal in the P-path transform domain to obtain the second real number signal in the P-path transform domain.

[0022] In combination with the first aspect, in some implementations of the first aspect, the P-path real-number signal includes an in-phase real-number signal I in the first polarization direction. m x and the orthogonal real signal Q m x and the in-phase real number signal I in the second polarization direction m y and the orthogonal real signal Q m y In the case where additional dispersion compensation is required, the dispersion compensation process includes: the dispersion impulse response I corresponding to the in-phase real signal h (t) and the dispersion impulse response Q corresponding to the orthogonal real signal h (t), perform NTT processing to obtain the dispersion equalization function I in the transform domain corresponding to the in-phase real signal h(w) and the dispersion equalization function Q in the transform domain corresponding to the orthogonal real signal h (w).

[0023] After obtaining the above I h (w) and Q h (w) After that, it is possible to transform the first real number signal in the P-way domain based on the in-phase real number signal, the orthogonal real number signal, the I h (w) and the Q h (w), determining the third real number signal of the 2*m-way transform domain in the first polarization direction in the third real number signal of the P-way transform domain, and the in-phase real number signal, the orthogonal real number signal, the I h (w) and the Q h (w) Determine that the third real signal in the 2*m-way transform domain in the second polarization direction in the third real signal in the P-way transform domain completes dispersion compensation.

[0024] In conjunction with the first aspect, in certain implementations of the first aspect, the third real-number signal in the transform domain after dispersion compensation and the first real-number signal in the transform domain meet the following requirements:

[0025] I' x (w)=I x (w)·I h (w)-Q x (w)·Q h (w);

[0026] Q' x (w) = Q x (w)·I h (w)+I x (w)·Q h (w);

[0027] I' y (w)=I y (w)·I h (w)-Q y (w)·Q h (w);

[0028] Q' y (w) = Q y (w)·I h (w)+I y (w)·Q h (w);

[0029] Among them, I x (w) represents the in-phase real signal in the first polarization direction in the first real signal of the transform domain, Q x(w) represents the orthogonal real signal in the first polarization direction in the first real signal of the transform domain, I y (w) represents the in-phase real signal in the second polarization direction in the first real signal of the transform domain, Q y (w) represents the orthogonal real signal in the second polarization direction in the first real signal of the transform domain, I' x (w) represents the in-phase real signal in the first polarization direction in the third real signal of the transform domain, Q' x (w) represents the orthogonal real signal in the first polarization direction in the third real signal of the transform domain, I' y (w) represents the in-phase real signal in the second polarization direction in the third real signal of the transform domain, Q' y (w) represents the orthogonal real signal in the second polarization direction in the third real signal in the transform domain.

[0030] Specifically, when m is equal to 1, based on the in-phase real signal, the orthogonal real signal, the I h (w) and the Q h (w) determining the third real signal in the m-path transform domain in the first polarization direction in the third real signal in the P-path transform domain includes:

[0031] Ix_3=Ix_1·I h (w)-Qx_1·Q h (w), Qx_3=Qx_1·I h (w)+Ix_1·Q h (w),

[0032] Wherein, Ix_3 represents the third real signal of the in-phase transform domain in the first polarization direction, Ix_1 represents the in-phase real signal in the first polarization direction in the first real signal of the P-path transform domain, Qx_3 represents the third real signal of the in-phase transform domain in the first polarization direction, and Qx_1 represents the in-phase real signal in the first polarization direction in the first real signal of the P-path transform domain;

[0033] The in-phase real number signal, the orthogonal real number signal, the I h (w) and the Q h (w) determining the third real signal in the 2*m-way transform domain in the second polarization direction in the third real signal in the P-way transform domain includes:

[0034] Iy_3=Iy_1·I h (w)-Qy_1·Q h (w), Qy_3=Qy_1·Ih (w)+Iy_1·Q h (w),

[0035] The Iy_3 represents the third real signal of the in-phase transformation domain in the second polarization direction, the Iy_1 represents the in-phase real signal in the second polarization direction in the first real signal of the P-path transformation domain, the Qy_3 represents the third real signal of the in-phase transformation domain in the second polarization direction, and the Qy_1 represents the in-phase real signal in the second polarization direction in the first real signal of the P-path transformation domain.

[0036] In combination with the first aspect, in certain implementations of the first aspect, the above-mentioned DBP processing specifically includes: the P-path real number signal is subjected to NTT processing to obtain a fourth real number signal in the P-path transform domain; the fourth real number signal in the P-path transform domain is subjected to dispersion compensation in the transform domain to obtain a fifth real number signal in the P-path transform domain; the fifth real number signal in the P-path transform domain is subjected to INTT processing to obtain a ninth real number signal in the P-path time domain; the ninth real number signal in the P-path time domain is subjected to nonlinear compensation to obtain a tenth real number signal in the P-path time domain.

[0037] In combination with the first aspect, in certain implementations of the first aspect, polarization compensation may be performed first and then INTT processing may be performed to obtain a seventh real signal in the P-channel time domain.

[0038] Specifically, the second real-number signal in the P-channel transform domain is polarization compensated to obtain a sixth real-number signal in the P-channel transform domain; the sixth real-number signal in the P-channel transform domain is INTT-processed to obtain a seventh real-number signal in the P-channel time domain. The second real-number signal in the 2*m-channel transform domain in the first polarization direction within the second real-number signal in the P-channel transform domain is equalized and depolarized to obtain a sixth real-number signal in the 2*m-channel transform domain in the first polarization direction within the sixth real-number signal in the P-channel transform domain; the second real-number signal in the 2*m-channel transform domain in the second polarization direction within the second real-number signal in the P-channel transform domain is equalized and depolarized to obtain a sixth real-number signal in the 2*m-channel transform domain in the second polarization direction within the sixth real-number signal in the P-channel transform domain.

[0039] Then, the seventh real signal of the P-path time domain obtained by INTT processing is merged, wherein the 2*m-path seventh real signal of the P-path time domain in the first polarization direction is merged into the m-path complex signal X in the time domain in the first polarization direction, and the 2*m-path seventh real signal of the P-path time domain in the second polarization direction is merged into the m-path complex signal Y in the time domain in the second polarization direction.

[0040] In combination with the first aspect, in some implementations of the first aspect, INTT processing may be performed first and then polarization compensation may be performed to obtain a time-domain complex signal X and a time-domain complex signal Y after polarization compensation.

[0041] Specifically, the second real number signal in the P-path transform domain is processed by INTT to obtain an eighth real number signal in the P-path time domain, wherein the 2*m-path eighth real number signals in the first polarization direction in the eighth real number signal in the P-path time domain are merged into the m-path complex signal X in the time domain in the first polarization direction, and the 2*m-path eighth real number signals in the second polarization direction in the eighth real number signal in the P-path time domain are merged into the m-path complex signal Y in the time domain in the second polarization direction; the m-path complex signal X in the time domain in the first polarization direction and the m-path complex signal Y in the time domain in the second polarization direction are polarization compensated in the time domain to obtain the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain after polarization compensation.

[0042] In conjunction with the first aspect, in certain implementations of the first aspect, the P real-number signals include 4*m real-number signals in an m transmission mode. m can be a value equal to or greater than 1, and P and m satisfy P=4*m. When m is equal to 1, it can be understood as single-mode transmission; when m is greater than 1, it can be understood as multi-mode transmission.

[0043] In a second aspect, a signal processing apparatus is provided, comprising a processor coupled to a memory and configured to execute instructions in the memory to implement the method in the first aspect and any possible implementation of the first aspect.

[0044] Optionally, the device further includes a communication interface, and the processor is coupled to the communication interface.

[0045] In one implementation, the device is a digital signal processor. When the device is a digital signal processor, the communication interface may be a transceiver, or an input / output interface.

[0046] In another implementation, the device is a chip configured in a digital signal processor. When the device is a chip configured in a digital signal processor, the communication interface may be an input / output interface.

[0047] In another implementation, the device is a chip or a chip system.

[0048] Optionally, the transceiver may be a transceiver circuit.

[0049] Optionally, the input / output interface may be an input / output circuit.

[0050] Specifically, the signal processing device includes:

[0051] A first number theory transform (NTT) module is configured to perform NTT processing on the P-channel input signal to obtain a first real number signal in a P-channel transform domain;

[0052] a clock recovery module, configured to perform clock recovery on the first real number signal in the P-path transform domain, or the first real number signal in the P-path transform domain after dispersion compensation, to obtain a second real number signal in the P-path transform domain;

[0053] The polarization compensation module and the first inverse number theory transformation INTT module are used to process the second real number signal in the P-channel transformation domain to obtain m-channel complex signals X in the time domain and m-channel complex signals Y in the time domain;

[0054] A phase recovery module and a decoding module, configured to perform phase recovery and decoding on the complex signal X in the time domain and the complex signal Y in the time domain to obtain a bit signal;

[0055] Wherein, m and P are positive integers.

[0056] In combination with the second aspect, in certain implementations of the second aspect, the P-path input signal includes a P-path real number signal, and the first number theoretic transform NTT module, which is used to perform NTT processing on the P-path input signal to obtain a first real number signal in the P-path transform domain, includes: the first number theoretic transform NTT module, which is used to perform number theoretic transform on the P-path real number signal to obtain a first real number signal in the P-path transform domain.

[0057] For example, the P-channel real-number signal includes 4*m-channel real-number signals in the m transmission mode. m can be a value equal to or greater than 1, and P and m satisfy P=4*m. When m is equal to 1, it can be understood as single-mode transmission; when m is greater than 1, it can be understood as multi-mode transmission.

[0058] In combination with the second aspect, in certain implementations of the second aspect, the device also includes: a first dispersion compensation module, used to perform dispersion compensation on the first real number signal in the P-path transform domain to obtain a third real number signal in the P-path transform domain; and a clock recovery module, used to perform clock recovery on the third real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain.

[0059] In conjunction with the second aspect, in certain implementations of the second aspect, the first dispersion compensation module, configured to perform dispersion compensation on the first real signal in the P-path transform domain to obtain a third real signal in the P-path transform domain, includes:

[0060] The first dispersion compensation module is used to calculate the dispersion equalization function I of the transform domain corresponding to the in-phase real number signal, the orthogonal real number signal, and the in-phase real number signal in the first polarization direction in the first real number signal of the P-path transform domain. h(w) and the dispersion equalization function Q in the transform domain corresponding to the orthogonal real signal h (w) determining a third real signal in the m-way transform domain in the first polarization direction in the third real signal in the P-way transform domain;

[0061] The first dispersion compensation module is used to convert the first real number signal in the P-path transform domain into an in-phase real number signal in the second polarization direction, an orthogonal real number signal, and the I h (w) and the Q h (w), determining a third real number signal of the 2*m-way transform domain in the second polarization direction in the third real number signal of the P-way transform domain, wherein the I h (w) and the Q h (w) are the dispersion impulse responses I corresponding to the in-phase real signal h (t) and the dispersion impulse response Q corresponding to the orthogonal real signal h (t)NTT processing is obtained.

[0062] In conjunction with the second aspect, in certain implementations of the second aspect, the third real-number signal in the transform domain after dispersion compensation and the first real-number signal in the transform domain meet the following requirements:

[0063] I' x (w)=I x (w)·I h (w)-Q x (w)·Q h (w);

[0064] Q' x (w) = Q x (w)·I h (w)+I x (w)·Q h (w);

[0065] I' y (w)=I y (w)·I h (w)-Q y (w)·Q h (w);

[0066] Q' y (w) = Q y (w)·I h (w)+I y (w)·Q h (w);

[0067] Among them, I x (w) represents the in-phase real signal in the first polarization direction in the first real signal of the transform domain, Q x(w) represents the orthogonal real signal in the first polarization direction in the first real signal of the transform domain, I y (w) represents the in-phase real signal in the second polarization direction in the first real signal of the transform domain, Q y (w) represents the orthogonal real signal in the second polarization direction in the first real signal of the transform domain, I' x (w) represents the in-phase real signal in the first polarization direction in the third real signal of the transform domain, Q' x (w) represents the orthogonal real signal in the first polarization direction in the third real signal of the transform domain, I' y (w) represents the in-phase real signal in the second polarization direction in the third real signal of the transform domain, Q' y (w) represents the orthogonal real signal in the second polarization direction in the third real signal in the transform domain.

[0068] In combination with the second aspect, in certain implementations of the second aspect, the first dispersion compensation module includes a third NTT module, a merging module, and a multiplication module, wherein the third NTT module is used to perform NTT processing on the dispersion time domain impulse response to obtain a dispersion equalization function and to perform NTT processing on the first real number signal in the P-path transform domain, the multiplication module is used to multiply the first real number signal in the P-path transform domain after the NTT processing with the dispersion equalization function, and the merging module is used to merge the signals processed by the multiplication module.

[0069] In combination with the second aspect, in certain implementations of the second aspect, the P-channel input signal includes a P-channel tenth real number signal in the time domain; the device also includes: a digital back propagation DBP module, used to perform DBP processing on the P-channel digital signal to obtain the P-channel tenth real number signal in the time domain.

[0070] In the case where the device includes a DBP module, the device utilizes the DBP module to compensate for nonlinear effects on the basis of linear damage, thereby increasing the signal transmission distance.

[0071] In conjunction with the second aspect, in certain implementations of the second aspect, the DBP module includes, in sequence:

[0072] a second NTT module, a second dispersion compensation module, a second INTT module, and a nonlinear compensation module;

[0073] The DBP module is used to perform DBP processing on the P-channel digital signal to obtain the P-channel time-domain tenth real number signal, including:

[0074] The second NTT module is used to perform NTT processing on the P-channel digital signal to obtain a fourth real number signal in the P-channel transform domain;

[0075] The second dispersion compensation module is used to perform dispersion compensation on the fourth real number signal in the P-path transform domain in the transform domain to obtain a fifth real number signal in the P-path transform domain;

[0076] The second INTT module is used to perform INTT processing on the fifth real number signal in the P-channel transform domain to obtain a ninth real number signal in the P-channel time domain;

[0077] The nonlinear compensation module is used to perform nonlinear compensation on the ninth real-number signal in the P-channel time domain to obtain a tenth real-number signal in the P-channel time domain.

[0078] In combination with the second aspect, in some implementations of the second aspect, the clock recovery module is configured to perform clock recovery on the first real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain.

[0079] In conjunction with the second aspect, in some implementations of the second aspect, the apparatus further includes:

[0080] Merge modules;

[0081] The polarization compensation module is used to perform polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain;

[0082] The first inverse number theory transform INTT module is used to perform INTT processing on the sixth real number signal in the P-channel transform domain to obtain the seventh real number signal in the P-channel time domain;

[0083] The merging module is used to merge every two signals of the 2*m-way seventh real number signals in the time domain in the first polarization direction in the P-way seventh real number signal to obtain the m-way complex number signal X in the time domain in the first polarization direction, and

[0084] Each two signals of the 2*m-way seventh real number signals in the time domain in the second polarization direction in the P-way seventh real number signals in the time domain are combined to obtain the m-way complex number signals Y in the time domain in the second polarization direction.

[0085] Exemplarily, the polarization compensation module includes a first butterfly filter and a second butterfly filter;

[0086] The first butterfly filter is used to equalize and depolarize the second real number signal of the 2*m-way transform domain in the first polarization direction in the second real number signal of the P-way transform domain to obtain the sixth real number signal of the 2*m-way transform domain in the first polarization direction in the sixth real number signal of the P-way transform domain;

[0087] The second butterfly filter is used to equalize and depolarize the second real number signal of the 2*m-way transform domain in the second polarization direction in the second real number signal of the P-way transform domain to obtain the sixth real number signal of the 2*m-way transform domain in the second polarization direction in the sixth real number signal of the P-way transform domain.

[0088] Specifically, when the device processes signals, polarization compensation is first performed by the polarization compensation module and then by the first INTT module. Polarization compensation is performed on the signal in the transform domain. Performing polarization equalization damage compensation in the transform domain, replacing convolution with multiplication, can reduce power consumption.

[0089] or,

[0090] The apparatus further includes: a merging module; the first inverse number theory transform INTT module, configured to perform INTT processing on the second real number signal in the P-path transform domain to obtain an eighth real number signal in the P-path time domain;

[0091] The merging module is used to merge every two signals of the 2*m-way eighth real-number signals in the time domain in the first polarization direction in the P-way eighth real-number signals to obtain the m-way complex signal X in the time domain in the first polarization direction, and

[0092] Merging every two of the 2*m-path eighth real-number signals in the time domain in the second polarization direction in the P-path eighth real-number signals to obtain an m-path complex signal Y in the time domain in the second polarization direction;

[0093] The polarization compensation module is used for performing time-domain polarization compensation on m-channel complex signals X and m-channel complex signals Y.

[0094] Exemplarily, the polarization compensation module includes a third butterfly filter; the third butterfly filter is used to perform time-domain polarization compensation on the m-path complex signal X in the time domain in the first polarization direction and the m-path complex signal Y in the time domain in the second polarization direction to obtain the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain after polarization compensation.

[0095] Specifically, when the above-mentioned device processes signals, it first performs INTT processing based on the first INTT module and then performs polarization compensation processing based on the polarization compensation module. The polarization compensation processing is performed based on the time domain signal after INTT processing. Performing polarization compensation processing in the time domain can enhance the ability to resist loop delay.

[0096] According to a third aspect, a coherent receiver is provided, which includes the signal processing device according to the second aspect and any possible implementation of the second aspect.

[0097] Furthermore, the coherent receiver also includes a polarization beam splitter, a mixer, a photodetector and an analog-to-digital converter, wherein the polarization beam splitter is used to obtain signals in two polarization directions, the mixer is used to mix signals in the same polarization direction, the photodetector is used to convert the intensity of the optical signal into the intensity of the electrical signal, and the analog-to-digital converter is used to convert the signal from an analog signal to a digital signal.

[0098] In a fourth aspect, a chip is provided, including a communication interface, a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to read and execute the computer program stored in the memory, so that the chip implements the signal processing method in the above-mentioned first aspect and any possible implementation of the first aspect.

[0099] In a fifth aspect, a signal processing device is provided, comprising a processor and a communication interface, wherein the processor is coupled to a memory through the interface circuit, and the processor is used to execute program code in the memory to implement the signal processing method in the above-mentioned first aspect and any possible implementation of the first aspect.

[0100] In a sixth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a device, the device implements the signal processing method in the above-mentioned first aspect and any possible implementation of the first aspect.

[0101] In a seventh aspect, a computer program product comprising instructions is provided, which, when executed by a computer, enables a device to implement the signal processing method in the above-mentioned first aspect and any possible implementation of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0102] Figure 1 Schematic diagram of a polarization-multiplexed coherent receiver.

[0103] Figure 2 This is a schematic diagram of an Rx DSP architecture provided in an embodiment of the present application.

[0104] Figure 3 This is another schematic diagram of the Rx DSP structure provided in an embodiment of the present application.

[0105] Figure 4 (a) is a schematic diagram of dispersion compensation based on FFT transformation; Figure 4 (b) is a schematic diagram of the dispersion compensation module structure provided in an embodiment of the present application.

[0106] Figure 5 This is another Rx DSP structure diagram provided in an embodiment of the present application.

[0107] Figure 6 This is another Rx DSP structure diagram provided in an embodiment of the present application.

[0108] Figure 7 This is a structural diagram of a DBP module provided in an embodiment of the present application.

[0109] FIG8( a ) is a schematic structural diagram of a first polarization compensation module provided in an embodiment of the present application; FIG8( b ) is a schematic structural diagram of a second polarization compensation module provided in an embodiment of the present application.

[0110] Figure 9(a) is a schematic diagram of the Rx DSP architecture in a multi-mode transmission scenario provided in an embodiment of the present application; Figure 9(b) is a schematic diagram of the Rx DSP architecture in another multi-mode transmission scenario provided in an embodiment of the present application; Figure 9(c) is a schematic diagram of the Rx DSP architecture in yet another multi-mode transmission scenario provided in an embodiment of the present application; Figure 9(d) is a schematic diagram of the Rx DSP architecture in yet another multi-mode transmission scenario provided in an embodiment of the present application; Figure 9(e) is a schematic structural diagram of another DBP module provided in an embodiment of the present application.

[0111] Figure 10(a) is a schematic flowchart of a signal processing method provided in an embodiment of the present application; Figure 10(b) is a schematic flowchart of DBP processing provided in an embodiment of the present application; Figure 10(c) is a schematic flowchart of a method for obtaining a time domain complex signal provided in an embodiment of the present application; Figure 10(d) is another schematic flowchart of a method for obtaining a time domain complex signal provided in an embodiment of the present application.

[0112] Figure 11 This is a schematic diagram of a chip provided in an embodiment of the present application. DETAILED DESCRIPTION

[0113] The technical solution in this application will be described below with reference to the accompanying drawings.

[0114] The technical solution of the embodiment of the present application relates to the improvement of the polarization multiplexing coherent receiver. Figure 1 Explain the traditional polarization multiplexing coherent receiver, Figure 1 Schematic diagram of a polarization-multiplexed single-mode coherent receiver.

[0115] from Figure 1 It can be seen that the signal processing process of the polarization multiplexing single-mode coherent receiver includes:

[0116] The optical signal is processed by polarization beam splitter #1 in polarization beam splitter 101 and split into x1 and y1 signals, wherein x1 is sent to mixer #1 in mixer 102 and y1 is sent to mixer #2 in mixer 102.

[0117] The local oscillator light (e.g., generated by a local laser or a transmitter) is processed by the polarization beam splitter #2 in the polarization beam splitter 101 and is divided into x2 and y2 signals, where x2 is sent to the mixer #1 in the mixer 102 and y2 is sent to the mixer #2 in the mixer 102.

[0118] It should be noted that the optical signal or local oscillator light has two polarization modes, transverse electric (TE) and transverse magnetic (TM), or they can also be called X polarization state and Y polarization state, or they can also be called X polarization direction and Y polarization direction.

[0119] The X and Y polarization states are orthogonal to each other. That is, a light beam with a single polarization state (polarization state Y) undergoes polarization rotation and becomes a light beam with polarization state Y. It should be noted that the X and Y mentioned here do not refer to the narrow x and y coordinate axes, but rather to two mutually orthogonal directions, such as the horizontal and vertical directions of +45° and -45°.

[0120] Mixer #1 and Mixer #2 output four signals, which are input to photodiodes #1 to #4 as inputs to photodiodes 103. The photodiodes are used to convert the intensity of optical signals into the intensity of electrical signals.

[0121] The output signal of photodetector #1 is an Ix signal, the output signal of photodetector #2 is a Qx signal, the output signal of photodetector #3 is an Iy signal, and the output signal of photodetector #4 is a Qy signal.

[0122] The aforementioned Ix, Qx, Iy, and Qy are converted by analog-to-digital converters #1 through #4 in analog-to-digital converter (ADC) 104, respectively, to output the converted digital signals Ix, Qx, Iy, and Qy. Ix and Qx are combined by combining module #1 in combining module 105 to produce a complex signal X. Iy and Qy are combined by combining module #2 in combining module 105 to produce a complex signal Y. For example, the combining modules referred to in this application may also be referred to as adders.

[0123] The complex signal X is input to FFT module #1 in FFT module 106 and processed by FFT module #1 to obtain a frequency domain signal X after fast Fourier transform. The complex signal Y is input to FFT module #2 in FFT module 106 and processed by FFT module #2 to obtain a frequency domain signal Y after fast Fourier transform.

[0124] Frequency domain signal X is input to dispersion compensation module #1 in dispersion compensation (CDC) module 112, where dispersion compensation is performed in the frequency domain to obtain compensated frequency domain signal X. Frequency domain signal Y is input to dispersion compensation module #2 in dispersion compensation module 112, where dispersion compensation is performed in the frequency domain to obtain compensated frequency domain signal Y.

[0125] The frequency domain signal X after frequency domain dispersion compensation is input to the clock recovery module #1 in the clock recovery module 107 to complete clock recovery to obtain the frequency domain signal X after clock recovery. The frequency domain signal Y after frequency domain dispersion compensation is input to the clock recovery module #2 in the clock recovery module 107 to complete clock recovery to obtain the frequency domain signal Y after clock recovery.

[0126] The frequency domain signal X after clock recovery is input to IFFT module #1 in IFFT module 113 and processed by IFFT module #1 to obtain the time domain signal X after inverse fast Fourier transform. The frequency domain signal Y after clock recovery is input to IFFT module #2 and processed by IFFT module #2 in IFFT module 113 to obtain the time domain signal Y after inverse fast Fourier transform.

[0127] The time domain signal X and the time domain signal Y are simultaneously input into the polarization compensation module 108. Optionally, the polarization compensation module is composed of a 2×2 butterfly filter to complete polarization demultiplexing and damage equalization to obtain the time domain signal X and the time domain signal Y after polarization compensation. Specifically, Figure 1 The polarization compensation module 108 further includes a coefficient updating module 109 , which is used to update the coefficients of the filter included in the polarization compensation module 108 .

[0128] The polarization-compensated complex signal X is sequentially fed into phase recovery module #1 in carrier phase recovery module 110 and decoding module #1 in decoder module 111 to obtain a bit signal. The polarization-compensated complex signal Y is sequentially fed into phase recovery module #2 in phase recovery module 110 and decoding module #2 in decoding module 111 to obtain a bit signal.

[0129] It should be noted that the introduction of the FFT module 106 and the IFFT module 113 into the signal processing flow of the current coherent receiver may lead to the following situations:

[0130] 1) The complexity of FFT transform increases at a rate of Nlog2N with the number of transform points;

[0131] 2) FFT transformation involves a large number of multiplication operations, and the complexity of multiplication is very high;

[0132] 3) The transformation matrix of FFT is a complex matrix based on the e-index, which cannot be accurately represented in a computer with a finite number of bits and has truncation errors and fixed-point costs.

[0133] Since the embodiments of the present application mainly involve the improvement of the processing flow after ADC, for the convenience of description, the processing flow after ADC is collectively referred to as the receiver digital signal processing (Receiver Digital Signal Processing, Rx DSP) process, such as Figure 1 The large dashed box includes the Rx DSP signal processing process.

[0134] To facilitate understanding, several basic concepts involved in this application are briefly introduced:

[0135] 1. Polarization multiplexing.

[0136] The transmission mode of optical signal in single-mode fiber is HE 11 The two sub-patterns of the pattern HE x 11 、HE y 11 The two submodes are independent and have orthogonal polarization directions. During transmission, the two submodes may exhibit different forms of linear polarization, elliptical polarization, and circular polarization, but they always remain orthogonal. The X polarization and Y polarization involved in this application refer to the two orthogonal polarization states in a multiplexed single-mode optical fiber.

[0137] Compared with single-mode optical fiber, multimode optical fiber can transmit signals in multiple modes. The most commonly used mode is LP. pq The values ​​of mode, p, and q represent the different mode field characteristics of the LP mode.

[0138] 2. NTT processing and INTT processing.

[0139] The NTT processing and FFT processing involved in this application are both transformations that implement fast convolution. Among them, FFT processing is a complex transformation based on the core K = exp(jα). Exp(jα) is a complex number, and its storage has truncation errors. At the same time, the multiplication complexity of complex numbers is very high. The NTT transformation is a transformation defined on a finite field with a core of K (K is an integer). Generally, K is 2, so the transformation matrix has no truncation errors. In binary computers, multiplication and division by 2 can be implemented by shifting, which can eliminate multiplication and reduce power consumption.

[0140] In addition, INTT processing is the inverse process of the above-mentioned NTT processing, and IFFT is the inverse process of the above-mentioned FFT processing, which will not be described in detail in this application.

[0141] 3. Transform domain and time domain.

[0142] The time domain signal involved in the embodiments of the present application refers to the signal input into the Rx DSP process after analog-to-digital conversion. The signal obtained by FFT processing of the time domain signal is called a frequency domain signal, and the signal obtained by NTT processing is called a transform domain signal.

[0143] In addition, the frequency domain signal is processed by IFFT to obtain the time domain signal, and the transform domain signal is processed by INTT to obtain the time domain signal.

[0144] In other words, the time domain signal in this application can be understood as a signal at a time granularity that may change over time; the frequency domain signal can be understood as a time domain signal converted to a frequency domain representation by means of a time domain to frequency domain transformation. Similarly, the transform domain signal can be understood as a time domain signal converted to a transform domain representation by means of a time domain to transform domain transformation. In the process of transforming from the time domain to the transform domain, the time domain signal is usually segmented, and then a portion of the overlap is added before being fed into the NTT / INTT transformation module. For example, if a segment of time domain data is 1024 points, the time domain signal is first divided into 512 points, 256 points are then added, and the 1024 points are then fed into the NTT / INTT transformation module. The 512, 256, and 1024 here are just examples. The specific overlap ratio and the number of segmentation points are not limited in this application. The number of points in data transformation refers to the length of the array during data transformation. Generally speaking, the larger the number of points, the better the transformation performance.

[0145] 4. In-phase real signal and orthogonal real signal.

[0146] Quadrature amplitude modulation (QAM) is widely used in modern coherent communications. It is a modulation method that performs amplitude modulation on two orthogonal carrier waves. These two carrier waves are typically sinusoidal waves with a 90° phase difference, hence the name "quadrature carriers." The real part of this complex signal is called the in-phase component, while the imaginary part is called the quadrature component.

[0147] It should be noted that the real and imaginary parts of the complex signal (e.g., X=a+bj) are both real numbers (e.g., a and b are both real numbers), so in this application, the above-mentioned in-phase components are called in-phase real signals, and the orthogonal components are called orthogonal real signals.

[0148] From the above, it can be seen that the introduction of FFT modules and IFFT modules in the process of processing electrical signals in a coherent receiver may increase the complexity of the signal processing process. In order to solve the problems existing in the existing process of processing electrical signals in a coherent receiver and reduce the power consumption of the algorithm, an embodiment of the present application provides a signal processing method and device, which completes the processing of electrical signals by avoiding the introduction of the above-mentioned FFT modules and IFFT modules, thereby improving the performance of polarization multiplexing coherent receivers in processing electrical signals and reducing the complexity.

[0149] The signal processing method and apparatus provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings. The signal processing method provided by the embodiments of the present application can be applied to a polarization multiplexing coherent receiver to process electrical signals in single-mode or multi-mode transmission scenarios.

[0150] Figure 2 2 is a schematic diagram of an Rx DSP architecture provided in an embodiment of the present application. The Rx DSP architecture 200 includes:

[0151] A first NTT module 211 , a first dispersion compensation module 212 , a clock recovery module 213 , a first polarization compensation module 214 , a first INTT module 215 , a combining module 216 , a phase recovery module 217 and a decoding module 218 .

[0152] The first NTT module includes NTT module #1, NTT module #2, NTT module #3 and NTT module #4. Signal I 1 x , Q 1 x , I 1 y and Q 1 y Input to NTT module #1, NTT module #2, NTT module #3 and NTT module #4 respectively.

[0153] Optionally, signal I 1 x , Q 1 x , I 1 y and Q 1 y It is the real signal after ADC conversion.

[0154] NTT module #1, NTT module #2, NTT module #3 and NTT module #4 are respectively 1 x , Q 1 x , I 1 y and Q 1y The NTT processing is performed to output the first real-number signals Ix_1 , Qx_1 , Iy_1 and Qy_1 in the transform domain to the first dispersion compensation module 212 .

[0155] First dispersion compensation module 212 includes dispersion compensation module #1 and dispersion compensation module #2. Signals Ix_1 and Qx_1 are input to dispersion compensation module #1 for dispersion compensation, resulting in third real-number signals Ix_3 and Qx_3 in the transform domain. Signals Iy_1 and Qy_1 are input to dispersion compensation module #2 for dispersion compensation, resulting in third real-number signals Iy_3 and Qy_3 in the transform domain.

[0156] Clock recovery module 213 includes clock recovery module #1 and clock recovery module #2. Signals Ix_3 and Qx_3 are input to clock recovery module #1 for clock recovery, resulting in transform-domain second real-number signals Ix_2 and Qx_2. Signals Iy_3 and Qy_3 are input to clock recovery module #2 for clock recovery, resulting in transform-domain second real-number signals Iy_2 and Qy_2.

[0157] The second real-number signals Ix_2, Qx_2, Iy_2, and Qy_2 are input to the first polarization compensation module 214 for polarization demultiplexing and damage equalization to generate sixth real-number signals Ix_5, Qx_5, Iy_5, and Qy_5 in the transform domain. Optionally, the first polarization compensation module 214 further includes a coefficient update module 2141 for updating the coefficients of the filter included in the first polarization compensation module 214.

[0158] The first INTT module 215 includes INTT module #1, INTT module #2, INTT module #3, and INTT module #4. Signals Ix_5, Qx_5, Iy_5, and Qy_5 are input to INTT module #1, INTT module #2, INTT module #3, and INTT module #4, respectively.

[0159] INTT module #1, INTT module #2, INTT module #3 and INTT module #4 perform INTT processing on Ix_5, Qx_5, Iy_5 and Qy_5 respectively and output seventh real-number signals Ix_6, Qx_6, Iy_6 and Qy_6 in the time domain to the merging module 216.

[0160] The merging module 216 includes a merging module #1 and a merging module #2. Ix_6 and Qx_6 are input to the merging module #1 for merging to obtain a time domain complex signal X; Iy_6 and Qy_6 are input to the merging module #2 for merging to obtain a time domain complex signal Y.

[0161] Phase recovery module 217 includes phase recovery module #1 and phase recovery module #2, and decoding module 111 includes decoding module #1 and decoding module #2. The time-domain complex signal X is sequentially input into phase recovery module #1 and decoding module #1 to produce a recovered bit signal. The time-domain complex signal Y is sequentially input into phase recovery module #2 and decoding module #2 to produce a recovered bit signal.

[0162] Figure 2 The illustrated Rx DSP architecture 200 can be referred to as a full-transform-domain Rx DSP architecture. This architecture not only performs dispersion compensation based on NTT / INTT, but also polarization compensation based on NTT / INTT, i.e., it performs both dispersion and polarization compensation in the transform domain. Specifically, polarization equalization impairment compensation is performed in the transform domain, replacing convolution with multiplication. This further reduces power consumption while increasing the number of transform points (size) to effectively equalize impairments and improve performance.

[0163] Figure 3 FIG. 3 is a schematic diagram of another Rx DSP architecture provided in an embodiment of the present application. The Rx DSP architecture 300 includes:

[0164] The first NTT module 211 , the first dispersion compensation module 212 , the clock recovery module 213 , the second polarization compensation module 310 , the coefficient update module 311 , the first INTT module 215 , the merging module 216 , the phase recovery module 320 and the decoding module 330 .

[0165] The modules of the architecture 300 before clock recovery are the same as those of the architecture 200. The difference is that the signal output by the clock recovery module 213 is first processed by the INTT module 215 and merged by the merging module 216, and then input into the second polarization compensation module 310. The process before clock recovery will not be repeated.

[0166] The first INTT module 215 includes INTT module #1, INTT module #2, INTT module #3 and INTT module #4. The second real-number signals Ix_2, Qx_2, Iy_2 and Qy_2 are input to INTT module #1, INTT module #2, INTT module #3 and INTT module #4 respectively;

[0167] INTT module #1, INTT module #2, INTT module #3 and INTT module #4 perform INTT processing on Ix_2, Qx_2, Iy_2 and Qy_2 respectively and output eighth real-number signals Ix_7, Qx_7, Iy_7 and Qy_7 in the time domain to the merging module 216.

[0168] The merging module 216 includes a merging module #1 and a merging module #2. Ix_7 and Qx_7 are input to the merging module #1 for merging to obtain a time domain complex signal X; Iy_7 and Qy_7 are input to the merging module #2 for merging to obtain a time domain complex signal Y.

[0169] The time-domain complex signal X and the time-domain complex signal Y are input to the second polarization compensation module 310 for polarization demultiplexing and damage equalization, thereby obtaining the polarization-compensated time-domain complex signal X and the time-domain complex signal Y. The second polarization compensation module 310 also includes a coefficient update module 311 for updating the coefficients of the filter included in the second polarization compensation module 310.

[0170] Phase recovery module 320 includes phase recovery module #1 and phase recovery module #2, and decoding module 111 includes decoding module #1 and decoding module #2. The polarization-compensated time-domain complex signal X is sequentially input into phase recovery module #1 and decoding module #1 to produce a recovered bit signal. The polarization-compensated time-domain complex signal Y is sequentially input into phase recovery module #2 and decoding module #2 to produce a recovered bit signal.

[0171] Figure 3 The illustrated Rx DSP architecture 300 can be referred to as a combined time-domain and transform-domain Rx DSP architecture. This architecture utilizes NTT / INTT for dispersion compensation, while polarization compensation is still performed in the time domain. Specifically, performing dispersion compensation in the transform domain, a power-intensive module, effectively reduces overall power consumption. Retaining the polarization equalization module for compensation in the time domain enhances its loop delay mitigation capabilities.

[0172] Figure 4 (a) is a schematic diagram of the structure of a dispersion compensation module 400. The dispersion compensation module 400 can be applied to Figure 1 In the architecture shown, the dispersion compensation module includes: an FFT module 402 , an FFT module 404 , a merging module 401 , a merging module 403 and a multiplication module 405 .

[0173] When a signal is transmitted in an optical fiber, the effect of optical fiber dispersion on the signal can be described by impulse response. Time domain dispersion equalization uses finite impulse response (FIR) to compensate, and the coefficients of FIR are Where β2 represents the fiber's dispersion coefficient, z represents the distance the signal travels in the fiber, j represents an imaginary number, and t represents time. Because the x- and y-polarization signals experience the same dispersion impairment, the same h(t) function is used for equalization.

[0174] exist Figure 1 In the FFT-based dispersion compensation module in the signal processing flow shown, the input of the dispersion compensation module is a complex signal X(ω) in the frequency domain.

[0175] For example, I h (t) and Q h (t) is the dispersion time domain impulse response, I h (t) and Q h (t) is combined into a complex impulse response H(t) by the merging module 403. H(t) is input into the FFT module 404 to obtain the frequency domain equalization function H(ω). H(ω) and X(ω) are input into the multiplication module 405 for multiplication to obtain the output of the dispersion module Xout(ω)=H(ω)·X(ω);

[0176] Figure 4 (b) of FIG. 1 is a schematic diagram of the structure of a dispersion compensation module 410 according to an embodiment of the present application. The dispersion compensation module 410 can be applied to the aforementioned RX DSP architectures 200 and 300. The dispersion compensation module includes: NTT module 411, NTT module 412, NTT module 419, NTT module 420, merging module 414, merging module 418, and multiplication module 413, multiplication module 415, multiplication module 416, and multiplication module 417.

[0177] exist Figure 2 and Figure 3 In the signal processing flow shown, the NTT-based dispersion compensation module independently performs dispersion compensation on the signal in the x-polarization direction and the signal in the y-polarization direction. The dispersion compensation process is explained using the dispersion compensation in the x-polarization direction as an example (e.g. Figure 4 (as shown in (b) in the figure).

[0178] For example, the dispersion impulse response h(t) in the x-polarization direction is divided into Ih(t) and Qh(t), where

[0179] Ih(t) and Qh(t) are input into NTT module 419 and NTT module 420 respectively to obtain the dispersion equalization function I in NTT transform domain. h (w) and Q h The co-directional signal Ix(t) and the orthogonal signal Qx(t) in the x-polarization direction are input into the NTT module 411 and the NTT module 412 respectively to obtain the signal I in the transform domain. x (w) and Q x (w).

[0180] from Figure 4 In (b), we can see the same direction signal I in the transform domain x (w) and the same direction component I of the equilibrium functionh (w) The result obtained by multiplication of the input multiplication module 413 and the orthogonal signal Q in the transform domain x (w) and the quadrature component Q of the equalization function h The result obtained by multiplying (w) in the multiplication module 415 is combined by the merging module 414 to become Ix_out(w)=I x (w)·I h (w)-Q x (w)·Q h (w);

[0181] Q x (w) and I h (w) The result of the multiplication is input to the multiplication module 416, and I x (w) and Q h The result obtained by multiplying (w) in the multiplication module 417 is combined by the merging module 418 to form Qx_out(w)=Q x (w)·I h (w)-I x (w)·Q h (w).

[0182] Similarly, when performing dispersion compensation on the y polarization direction, the same direction signal I y (t) and the quadrature signal Q y (t) are input into the NTT module to obtain the signal I in the transform domain y (w) and Q y (w), the dispersion compensation process in the y polarization direction can be described by the formula:

[0183] Iy_out(w)=I y (w)·I h (w)-Q y (w)·Q h (w)

[0184] Qy_out(w)=Q y (w)·I h (w)+I y (w)·Q h (w).

[0185] In actual operation, Figure 4 The equalization function H(ω) in the Fourier transform domain in (a) and Figure 4 (b) The dispersion equalization function in the number theory transform domain I h (w) and Q h (w) only needs to be calculated once and stored in memory, and can be called repeatedly for different input signals.

[0186] For example, the difference between an NTT transformation and an FFT transformation is explained using the following examples:

[0187] The transformation matrices of FFT and NTT can be expressed as:

[0188]

[0189] The difference is the transform core of FFT The transformation core of NTT is α, which is a power of 2. Usually α is 2 (it can also be (or other values, which are not limited in this application). Taking α=2 as an example, each element in the transformation matrix is ​​still a power of 2, and multiplication by 2 or a power of 2 in binary operations can be accomplished by bit shifting.

[0190] Correspondingly, taking the kernel α=2 as an example, the transformation matrix H of NTT and the transformation matrix H of INTT are -1 They are:

[0191]

[0192] The internal working process of clock recovery module #1 and clock recovery module #2 is the same as Figure 4 Similar to (b), we only need to change the dispersion equalization function I h (w) and Q h (w) can be replaced with a function that balances the clock delay. The delay caused by the clock is expressed as h(t) = exp(-2πft0) ​​using the finite impulse response in the time domain. In the clock recovery module, I h (w) and Q h (w) can be replaced by cos(2πft0) ​​and the NTT transformation, which will not be repeated here.

[0193] Figure 5 FIG. 5 is another schematic diagram of an Rx DSP architecture provided in an embodiment of the present application. The Rx DSP architecture 500 includes:

[0194] A digital back propagation (DBP) module 510 , a first NTT module 211 , a clock recovery module 213 , a first polarization compensation module 214 , a first INTT module 215 , a merging module 216 , a phase recovery module 217 and a decoding module 218 .

[0195] The architecture of the Rx DSP architecture 500 after the first polarization compensation module 214 is the same as the aforementioned Rx DSP architecture 200 , except that the signal before polarization compensation has undergone DBP processing. Since the DBP module 510 has the function of dispersion compensation, the process after the DBP module does not require the dispersion compensation module 212.

[0196] Signal I 1 x , Q 1 x , I 1 y and Q 1 y The signals are input to the DBP module 510 , which performs DBP processing and outputs the tenth real-number signals Ix_4 , Qx_4 , Iy_4 and Qy_4 in the time domain to the first NTT module 211 .

[0197] The first NTT module 211 includes NTT module #1, NTT module #2, NTT module #3, and NTT module #4. Ix_4, Qx_4, Iy_4, and Qy_4 are input to NTT module #1, NTT module #2, NTT module #3, and NTT module #4, respectively.

[0198] NTT module #1, NTT module #2, NTT module #3 and NTT module #4 perform NTT processing on Ix_4, Qx_4, Iy_4 and Qy_4 respectively, and output the first real number signals Ix_1, Qx_1, Iy_1 and Qy_1 in the transform domain to the clock recovery module 213.

[0199] Clock recovery module 213 includes clock recovery module #1 and clock recovery module #2. Ix_1 and Qx_1 are input to clock recovery module #1 for clock recovery to obtain second real-number signals Ix_2 and Qx_2 in the transform domain. Iy_1 and Qy_1 are input to clock recovery module #2 for clock recovery to obtain second real-number signals Iy_2 and Qy_2 in the transform domain.

[0200] The process after the clock recovery module 213 refers to Figure 2 The embodiment shown.

[0201] Figure 5 The Rx DSP architecture 500 shown can be referred to as an Rx DSP architecture with full transform domain added to DBP. Figure 2 The Rx DSP architecture of the full transform domain shown is different in that the DBP module is added. Figure 2 The full-conversion-domain Rx DSP architecture shown only compensates for linear fiber impairments, such as dispersion compensation. The full-conversion-domain Rx DSP architecture with DBP modules not only compensates for linear impairments, but also utilizes DBP modules to compensate for nonlinear effects, thereby increasing signal transmission distance.

[0202] Figure 6 FIG. 6 is another schematic diagram of an Rx DSP architecture provided in an embodiment of the present application. The Rx DSP architecture 600 includes:

[0203] The DBP module 510 , the first NTT module 211 , the clock recovery module 213 , the second polarization compensation module 310 , the coefficient update module 311 , the first INTT module 215 , the merging module 216 , the phase recovery module 320 and the decoding module 330 .

[0204] The architecture of the Rx DSP architecture 600 after the second polarization compensation module 310 is the same as the Rx DSP architecture 300 described above; the architecture before the second polarization compensation module 310 is the same as the Rx DSP architecture 500 described above, that is, the received ADC converted signal is first subjected to DBP processing, and the process before DBP processing and clock recovery is referred to. Figure 5 In the embodiment shown, the process after the clock recovery module 213 is referred to Figure 3 The embodiment shown.

[0205] Figure 6 The Rx DSP architecture 600 shown can be referred to as a combined Rx DSP architecture of the time domain and transform domain with DBP. Figure 3 The difference between the Rx DSP architecture shown in the time domain and transform domain combination is the addition of the DBP module. Figure 3 The combined time-domain and transform-domain Rx DSP architecture shown here only compensates for linear fiber impairments, such as dispersion compensation. The combined time-domain and transform-domain Rx DSP architecture with the addition of a DBP module not only compensates for linear impairments but also utilizes the DBP module to compensate for nonlinear effects, thereby increasing signal transmission distance.

[0206] Figure 7 Schematic diagram of a DBP module provided in an embodiment of the present application. The DBP module 700 includes:

[0207] The second NTT module 711, the second dispersion compensation module 712, the second INTT module 713 and the nonlinear compensation module 714, wherein the structure of the second dispersion compensation module 712 is as above Figure 4 As shown in (b) in the figure, it is not described in detail here.

[0208] from Figure 7 As can be seen from the DBP module structure diagram shown, Figure 5 and Figure 6 As shown in I 1 x , Q 1 x , I 1 y and Q 1 y After DBP processing, the real signals Ix_4, Qx_4, Iy_4 and Qy_4 are obtained, which specifically include:

[0209] The second NTT module 711 includes NTT module #1, NTT module #2, NTT module #3 and NTT module #4. 1 x , Q 1 x , I 1 y and Q 1 y Input to NTT module #1, NTT module #2, NTT module #3 and NTT module #4 respectively.

[0210] NTT module #1, NTT module #2, NTT module #3 and NTT module #4 are respectively 1 x , Q 1 x , I 1 y and Q 1 y The NTT processing is performed to output the fourth real-number signals Ix_4′, Qx_4′, Iy_4′ and Qy_4′ in the transform domain to the second dispersion compensation module 712 .

[0211] Second dispersion compensation module 712 includes dispersion compensation module #1 and dispersion compensation module #2. Signals Ix_4' and Qx_4' are input to dispersion compensation module #1 for dispersion compensation in the transform domain, resulting in fifth real-number signals Ix_4" and Qx_4". Signals Iy_4' and Qy_4' are input to dispersion compensation module #2 for dispersion compensation in the transform domain, resulting in fifth real-number signals Iy_4" and Qy_4".

[0212] The second INTT module 713 includes INTT module #1, INTT module #2, INTT module #3 and INTT module #4. Ix_4", Qx_4", Iy_4" and Qy_4" are input to INTT module #1, INTT module #2, INTT module #3 and INTT module #4 respectively.

[0213] INTT module #1, INTT module #2, INTT module #3 and INTT module #4 perform INTT processing on Ix_4", Qx_4", Iy_4" and Qy_4", respectively, and output ninth real-number signals Ix_4'", Qx_4'", Iy_4' and Qy_4'' in the time domain to the nonlinear compensation module 714.

[0214] The nonlinear compensation module 714 includes nonlinear compensation module #1 and nonlinear compensation module #2. Ix_4' and Qx_4' are input to nonlinear compensation module #1 for nonlinear compensation to obtain Ix_4 and Qx_4; Iy_4' and Qy_4' are input to nonlinear compensation module #2 for nonlinear compensation to obtain Iy_4 and Qy_4.

[0215] It should be noted that the processing flow in the above-mentioned DBP module needs to be repeated Ns times. In order to ensure performance, the number of DBP iterations Ns also needs to be increased when the number of link spans increases. Different DBP schemes are not restricted here. Depending on the power consumption required by the scenario, Ns can be set to equal the total number of link spans or set to a single-step DBP. The DBP module involved in the embodiment of the present application avoids the use of FFT and IIFF modules, and replaces them with NTT and INTT modules. In this way, during the signal processing process of the DBP module, repeating Ns transformations will not cause error accumulation.

[0216] FIG8( a ) is a schematic structural diagram of a first polarization compensation module 214 provided in an embodiment of the present application; FIG8( b ) is a schematic structural diagram of a second polarization compensation module 310 provided in an embodiment of the present application.

[0217] As can be seen from Figure 8(a), the first polarization compensation module 214 involved in the embodiment of the present application includes a butterfly filter 810 and a butterfly filter 820, wherein the butterfly filter 810 includes filter 801, filter 802, filter 803, filter 804, and adder 805 and adder 806; the butterfly filter 820 includes filter 807, filter 808, filter 809, filter 8010, and adder 8011 and adder 8012.

[0218] The outputs of filters 801 and 807 are negated, and the outputs of 803 and 809 are also negated. These four signals are input to adder 805 to obtain the dispersion-compensated signal Ix_5. The outputs of filters 801, 807, 803, and 809 are also input to adder 805 to obtain the dispersion-compensated signal Qx_5. The outputs of filters 802 and 808 are negated, and the outputs of 804 and 8010 are also negated. These four signals are input to adder 806 to obtain the dispersion-compensated signal Iy_5. The outputs of filters 802, 808, 804, and 8010 are also negated, and these four signals are output to adder 805 to obtain the dispersion-compensated signal Qy_5.

[0219] The clock recovery module outputs four signals, Ix_2, Qx_2, Iy_2, and Qy_2, which are fed into the polarization compensation module. The polarization compensation module contains two butterfly filters. Each butterfly filter represents a 2x2, two-input, two-output system, with two complex inputs and two complex outputs. Because the NTT transform processes real signals, the polarization equalization section is a 4x4, four-input, four-output system.

[0220] The first butterfly filter 810 is for equalization and depolarization of the I-signal (including the x-polarized I-signal Ix_2 and the y-polarized I-signal Iy_2), and the transformation matrix is: The second butterfly filter 820 is for the Q-path signal (including the x-polarized Q-path signal Qx_2 and the y-polarized Q-path signal Qy_2), and the transformation matrix is

[0221] The outputs of the polarization compensation module are four real-number signals Ix_5, Qx_5, Iy_5, and Qy_5. The process of polarization compensation performed by the first polarization compensation module 214 is described in conjunction with FIG8(a):

[0222] The input signals of the first polarization compensation module 214 are four real signals Ix_2, Qx_2, Iy_2 and Qy_2: the in-direction signal Ix_2 in the x polarization direction, the orthogonal signal Qx_2 in the x polarization direction, the in-direction signal Iy_2 in the y polarization direction, and the orthogonal signal Qy_2 in the y polarization direction.

[0223] Specifically, Ix_2 and I hxx The multiplication result is input to adder 805, and Qx_2 and Q hxx The result of the multiplication is input into adder 805, Iy_2 and I hxy The multiplication result is input into adder 805, Qy_2 and Q hxy The multiplication result is negatively input to the adder 805, and the output of the adder 805 is the polarization-compensated in-direction signal Ix_5 in the x-polarization direction.

[0224] Ix_2 and Q hxx The result of the multiplication is input to adder 8011, and Qx_2 and I hxx The result of the multiplication is input to adder 8011, Iy_2 and Q hxy The result of the multiplication is input to adder 8011, Qy_2 and I hxy The multiplication result is input to the adder 8011, and the output of the adder 8011 is the orthogonal signal Qx_5 in the x-polarization direction after polarization compensation;

[0225] Ix_2 and I hyx The multiplication result is input to adder 806, and Qx_2 and Qhyx The result of the multiplication is input into adder 806, and Iy_2 and I hyy The multiplication result is input into adder 806, Qy_2 and Q hyy The negative result of the sum is input to the adder 806, and the output of the adder 806 is the polarization-compensated y-polarization direction in-phase signal Iy_5;

[0226] Ix_2 and Q hyx The result of the multiplication is input to adder 8012, and Qx_2 and I hyx The result of the multiplication is input to adder 8012, Iy_2 and Q hyy The result of the multiplication is input to adder 8012, Qy_2 and I hyy The multiplication result is input to the adder 8012, and the output of the adder 8012 is the orthogonal signal Qy_5 in the y polarization direction after polarization compensation.

[0227] The interaction between the outputs Ix_5, Qx_5, Iy_5, and Qy_5 of the first polarization compensation module 214 and the equalization matrix and the input signals Ix_2, Qx_2, Iy_2, and Qy_2 is described by the formula:

[0228]

[0229] Exemplarily, the first polarization compensation module 214 further includes a coefficient updating module 2141 (not shown in the figure), which is used to update the coefficients of the filter included in the first polarization compensation module 214. In the embodiment of the present application, the polarization compensation module equalization coefficient matrix and The updating methods include blind updating, updating with training, updating without decision feedback, or updating with decision feedback, etc. The embodiment of the present application does not limit the method of updating the polarization compensation module coefficients and will not be repeated here.

[0230] As can be seen from Figure 8(b), the second polarization compensation module 310 involved in the embodiment of the present application is composed of a 2×2 butterfly filter, and the butterfly filter includes filter 801, filter 802, filter 803, a fourth filter 804, and adders 805 and 806.

[0231] The functions of the second polarization compensation module 310 include equalization compensation (state of polarization, SOP), (differential group delay, DGD), residual dispersion and other channel fiber effects. The fiber effect can be described by a 2×2 damage matrix, and the inverse matrix of the damage matrix can be obtained by different algorithms. The tap coefficients of the time domain butterfly filter 108 are hxx, hxy, hyx and hyy. The X and Y signals are convolved with the tap coefficient matrices hxx, hxy, hyx and hyy to obtain the polarization damage compensated signals Xout and Yout. Common algorithms for calculating the tap coefficient matrix include the constant modulus algorithm (CMA) and the least mean square (LMS), as well as various modified versions of LMS or other forms such as data-assisted LMS.

[0232] The merging module 216 outputs a complex signal X in the X polarization direction, which is input to the filter 801 and the filter 803. The merging module 216 outputs a complex signal Y in the Y polarization direction, which is input to the filter 802 and the filter 804. The outputs of the filters 801 and 803 are input to the adder 805, and the outputs of the filters 802 and 804 are input to the adder 806.

[0233] In FIG8( b ), hxx, hxy, hyx, and hyy respectively represent coefficients of the filter 801 , the filter 802 , the filter 803 , and the filter 804 .

[0234] It should be noted that although the present application describes the butterfly filter using Figures 8(a) and 8(b) as examples, the butterfly filters shown in Figures 8(a) and 8(b) are only exemplary and do not constitute any limitation on the scope of protection of the present application. The butterfly filters involved in the embodiments of the present application should be understood in a broad sense, and any filter or filter combination that can realize the deconvolution function can be called a butterfly filter.

[0235] The Rx DSP provided in the embodiment of the present application may also be applied in a multi-mode transmission scenario. For example, in an m transmission mode, the Rx DSP receives P real-number signals.

[0236] Figure 9(a) is a schematic diagram of an Rx DSP architecture in a multi-mode transmission scenario provided by an embodiment of the present application. The Rx DSP architecture 900 includes:

[0237] A first NTT module 211 , a first dispersion compensation module 212 , a clock recovery module 213 , a first polarization compensation module 214 , a first INTT module 215 , a combining module 216 , a phase recovery module 217 and a decoding module 218 .

[0238] P real-number signals are input to the first NTT module 211 , which performs NTT processing on the P real-number signals and outputs P first real-number signals in the transform domain to the first dispersion compensation module 212 .

[0239] Optionally, the P-channel real-number signals are real-number signals converted by an ADC, and the P-channel real-number signals include 4*m-channel real-number signals in an m transmission mode.

[0240] The first dispersion compensation module 212 performs dispersion compensation on the first real number signals in the P-path transform domain and outputs a third real number signal in the P-path transform domain to the clock recovery module 213 .

[0241] The clock recovery module 213 performs clock recovery on the third real number signal in the P-path transform domain and outputs a second real number signal in the P-path transform domain to the first polarization compensation module 214 .

[0242] The first polarization compensation module 214 performs polarization compensation processing on the second real number signal in the P-path transform domain and outputs a sixth real number signal in the P-path transform domain to the first INTT module 215 .

[0243] The first INTT module 215 performs INTT processing on the P-channel sixth real-number signals in the transform domain and outputs P-channel seventh real-number signals in the time domain to the merging module 216 .

[0244] The merging module 216 performs merging processing on the P-channel seventh real-number signals in the time domain and outputs m-channel complex-number signals X and m-channel complex-number signals Y to the phase recovery module 217 and the decoding module 218 .

[0245] The phase recovery module 217 and the decoding module 218 perform phase recovery and decoding processing on the m-path complex signal X and the m-path complex signal Y to obtain recovered bit signals.

[0246] FIG9( b ) is a schematic diagram of an Rx DSP architecture in another multi-mode transmission scenario provided by an embodiment of the present application. The Rx DSP architecture 910 includes:

[0247] The first NTT module 211 , the first dispersion compensation module 212 , the clock recovery module 213 , the second polarization compensation module 108 , the first INTT module 215 , the combining module 216 , the phase recovery module 320 and the decoding module 330 .

[0248] The modules of architecture 910 before clock recovery are the same as those of architecture 900. The difference is that the signal output by the clock recovery module is first processed by the INTT module and merged by the merging module, and then input into the second polarization compensation module 108. The process before clock recovery will not be repeated.

[0249] The first INTT module 215 performs INTT processing on the P-channel transform domain second real number signals and outputs P-channel time domain eighth real number signals to the merging module 216. The merging module 216 performs merging processing on the P-channel time domain eighth real number signals and outputs m-channel complex number signals X and m-channel complex number signals Y to the second polarization compensation module 310.

[0250] The second polarization compensation module 310 performs polarization compensation on the m-path complex signal X and the m-path complex signal Y, and outputs the polarization-compensated m-path complex signal X and the m-path complex signal Y in the time domain to the phase recovery module 320 and the decoding module 330. The second polarization compensation module 310 further includes a coefficient updating module 311 (not shown), which is configured to update the coefficients of the filter included in the second polarization compensation module 310.

[0251] The phase recovery module 320 and the decoding module 330 perform phase recovery and decoding processing on the m-path complex signal X and the m-path complex signal Y to obtain recovered bit signals.

[0252] FIG9( c ) is a schematic diagram of an Rx DSP architecture in another multi-mode transmission scenario provided by an embodiment of the present application. The Rx DSP architecture 920 includes:

[0253] DBP module 510 , first NTT module 211 , clock recovery module 213 , first polarization compensation module 214 , first INTT module 215 , merging module 216 , phase recovery module 217 and decoding module 218 .

[0254] The architecture of the Rx DSP architecture 920 after the first polarization compensation module 214 is the same as the aforementioned Rx DSP architecture 900 , except that the signal before polarization compensation has undergone DBP processing. Since the DBP module 510 has the function of dispersion compensation, the process after the DBP module does not require the dispersion compensation module 212.

[0255] The P-channel real-number signals are input to the DBP module 510 , which performs DBP processing and outputs the P-channel time-domain tenth real-number signals to the first NTT module 211 .

[0256] The first NTT module 211 performs NTT processing on the tenth real number signal in the P-channel time domain and outputs a first real number signal in the P-channel transform domain to the clock recovery module 213 .

[0257] The clock recovery module 213 performs clock recovery processing on the first real number signals in the P-path transform domain and outputs the second real number signals in the P-path transform domain to the first polarization compensation module 214 .

[0258] The process after the clock recovery module 213 refers to the embodiment shown in FIG9( a ).

[0259] FIG9( d ) is a schematic diagram of an Rx DSP architecture in another multi-mode transmission scenario provided by an embodiment of the present application. The Rx DSP architecture 930 includes:

[0260] The DBP module 510 , the first NTT module 211 , the clock recovery module 213 , the second polarization compensation module 310 , the first INTT module 215 , the merging module 216 , the phase recovery module 320 and the decoding module 330 .

[0261] The architecture of the Rx DSP architecture 930 after the second polarization compensation module 310 is the same as the above-mentioned Rx DSP architecture 910; the architecture before the second polarization compensation module 310 is the same as the above-mentioned Rx DSP architecture 920, that is, the received signal after ADC conversion is first subjected to DBP processing. The process before DBP processing and clock recovery refers to the embodiment shown in Figure 9(c), and the process after the clock recovery module 213 refers to the embodiment shown in Figure 9(b).

[0262] For example, the DBP module 510 shown in FIG. 9( c ) and FIG. 9( d ) is shown in FIG. 9( e ), which is a schematic structural diagram of another DBP module provided in an embodiment of the present application. The DBP module 940 includes:

[0263] A second NTT module 711 , a second dispersion compensation module 712 , a second INTT module 713 and a nonlinear compensation module 714 .

[0264] The P-channel real-number signal is input to the DBP module 510, and the P-channel time-domain tenth real-number signal is obtained through DBP processing. Specifically, the tenth real-number signal includes:

[0265] P real-number signals are respectively input into P second NTT modules 711 , which are processed by NTT and output as P fourth real-number signals in the transform domain to the second dispersion compensation module 712 .

[0266] The second dispersion compensation module 712 performs dispersion compensation processing on the fourth real-number signal in the P-path transform domain and outputs a fifth real-number signal in the P-path transform domain to the second INTT module 713 .

[0267] The second INTT module 713 performs INTT processing on the fifth real-number signal in the P-channel transform domain and outputs a ninth real-number signal in the P-channel time domain to the nonlinear compensation module 714 .

[0268] The nonlinear compensation module 714 performs nonlinear compensation on the ninth real-number signal in the P-channel time domain to obtain a tenth real-number signal in the P-channel time domain.

[0269] Specifically, using the NTT module and INTT module for signal processing has the following advantages over using the FFT module and IFFT module for signal processing:

[0270] 1) Signal processing is performed based on the NTT module and the INTT module, reducing the complexity of the Rx DSP architecture.

[0271] Because signal processing is based on FFT and IFFT modules, dispersion compensation alone accounts for half of the total Rx DSP power consumption during signal transmission. The FFT complexity of N points is proportional to Nlog2N multiplications. Depending on the data bit width, the complexity of a multiplication is several to dozens of times that of an addition.

[0272] The signal processing complexity of NTT and INTT modules is proportional to 2Nlog2N additions. The complexity reduction achieved by these modules varies depending on the data bit width and the number of transformation points, N. For example, with 256 transformation points, NTT reduces power consumption to approximately one-quarter that of FFT.

[0273] 2) Based on the NTT module and the INTT module for signal processing, the transformation matrix has no truncation error and fixed-point cost.

[0274] Signal processing is performed based on the FFT module and IFFT module. Because the transformation core of FFT is an exponential function, it cannot be accurately expressed in binary computers and has truncation errors. At the same time, the transformation matrix bit width is uncertain and there is a fixed-point cost.

[0275] In signal processing based on the NTT module and the INTT module, because the transformation core of the NTT is an integer, the transformation core is generally 2 or a power of 2, which can be accurately expressed in the memory and the transformation matrix bit width is fixed, so there is no fixed-point cost.

[0276] 3) There is no error accumulation in signal processing based on the NTT module and the INTT module.

[0277] The truncation error and fixed-point cost in a single FFT or IFFT transform are repeatedly accumulated during DBP propagation, resulting in a larger error in the final result.

[0278] However, there is no truncation error or fixed-point cost in a single NTT or INTT transformation, so repeating the transformation Ns times will not cause error accumulation.

[0279] 4) Signal processing based on the NTT module and the INTT module can reduce storage.

[0280] The core of the NTT transform is 2, and the transform matrix contains 2 or powers of 2. Because multiplying a binary number by a power of 2 only requires shifting, there is no need to store a large transform matrix in the DSP process, only the core of the transform needs to be stored.

[0281] FIG10( a ) is a schematic flowchart of a signal processing method provided in an embodiment of the present application, which includes the following steps S1010 to S1050 .

[0282] S1010, obtain P real-number signals.

[0283] Exemplarily, P real-number signals include real-number signals in two polarization directions corresponding to each mode in m modes, where m equals 1 and represents single-mode transmission, and m greater than 1 represents multi-mode transmission, and m and P are positive integers.

[0284] S1020: Obtain a first real number signal in a P-path transform domain.

[0285] Perform at least NTT processing on the P-way real-number signal to obtain a first real-number signal in the P-way transform domain. There are two possible implementation methods:

[0286] Implementation 1.1: Perform NTT processing on P-channel real-number signals to obtain a first real-number signal in a P-channel transform domain.

[0287] Implementation 1.2: Perform DBP processing and NTT processing on P-channel real-number signals in sequence to obtain a first real-number signal in the P-channel transform domain.

[0288] In implementation 1.2, first, DBP processing is performed on the P-channel real-number signals to obtain the tenth real-number signal in the P-channel time domain, and then NTT processing is performed on the tenth real-number signal in the P-channel time domain to obtain the first real-number signal in the P-channel transform domain.

[0289] The above-mentioned implementation method 1.1 performs NTT processing on the P-path real number signal and implementation method 1.2 performs NTT processing on the tenth real number signal in the P-path time domain after the P-path real number signal is DBP processed. They can be collectively referred to as performing NTT processing on the P-path input signal, that is, in the embodiment of the present application, the P-path input signal includes the P-path real number signal, or the tenth real number signal in the P-path time domain after the P-path real number signal is DBP processed.

[0290] For ease of understanding, the process of performing DBP processing on P-channel real-number signals to obtain the tenth real-number signal in the P-channel time domain is briefly described in conjunction with FIG10( b ). FIG10( b ) is a schematic flow chart of the DBP processing provided in an embodiment of the present application, including the following steps S1021 to S1024:

[0291] S1021: Obtain a fourth real number signal in the P-way transform domain.

[0292] The P-way real-number signal is subjected to NTT processing to obtain a fourth real-number signal in the P-way transform domain;

[0293] S1022: Obtain a fifth real number signal in the P-way transform domain.

[0294] The fourth real-number signal in the P-path transform domain is dispersion compensated in the transform domain to obtain a fifth real-number signal in the P-path transform domain;

[0295] S1023 , obtain a ninth real-number signal in the P-channel time domain.

[0296] The fifth real-number signal in the P-path transform domain is processed by INTT to obtain a ninth real-number signal in the P-path time domain;

[0297] S1024: Obtain a tenth real number signal in the P-channel time domain.

[0298] The ninth real number signal of the P-path time domain is subjected to nonlinear compensation to obtain the tenth real number signal of the P-path time domain

[0299] S1030: Obtain a second real number signal in a P-path transform domain.

[0300] Perform at least clock recovery processing on the first real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain. Corresponding to the two possibilities in the above-mentioned step S1020, there are two possible implementations:

[0301] Implementation 2.1 corresponds to implementation 1.1 in step S1020 above, and the processing flow includes:

[0302] First, dispersion compensation is performed on the first real signal in the P-path transform domain to obtain a third real signal in the P-path transform domain;

[0303] Then, clock recovery is performed on the third real-number signals in the P-path transform domain to obtain second real-number signals in the P-path transform domain.

[0304] Performing dispersion compensation on the first real number signal in the P-path transform domain to obtain a third real number signal in the P-path transform domain specifically includes:

[0305] First, determine the dispersion impulse response I corresponding to the in-phase real signal h (t) and the dispersion impulse response Q corresponding to the orthogonal real signalh (t), the dispersion equalization function I corresponding to the in-phase real signal is obtained by NTT h (w) and the dispersion equalization function Q in the transform domain corresponding to the orthogonal real signal h (w);

[0306] Then, based on the in-phase real number signal, the orthogonal real number signal, and the I h (w) and Q h (w), determining the third real number signal of the 2*m-way transform domain in the first polarization direction in the third real number signal of the P-way transform domain; and based on the in-phase real number signal, the orthogonal real number signal, and the I h (w) and Q h (w) Determine a third real number signal in the 2*m-way transform domain in the second polarization direction in the third real number signal in the P-way transform domain.

[0307] Specifically, the third real-number signal in the transform domain after dispersion compensation and the first real-number signal in the transform domain before dispersion compensation meet the following requirements:

[0308] I' x (w)=I x (w)·I h (w)-Q x (w)·Q h (w);

[0309] Q' x (w) = Q x (w)·I h (w)+I x (w)·Q h (w);

[0310] I' y (w)=I y (w)·I h (w)-Q y (w)·Q h (w);

[0311] Q' y (w) = Q y (w)·I h (w)+I y (w)·Q h (w);

[0312] Among them, I x (w) represents the in-phase real signal in the first polarization direction in the first real signal of the transform domain, Q x(w) represents the orthogonal real signal in the first polarization direction in the first real signal of the transform domain, I y (w) represents the in-phase real signal in the second polarization direction in the first real signal in the transform domain, Q y (w) represents the orthogonal real signal in the second polarization direction in the first real signal of the transform domain, I' x (w) represents the in-phase real signal in the first polarization direction in the third real signal of the transform domain, Q' x (w) represents the orthogonal real signal in the first polarization direction in the third real signal of the transform domain, I' y (w) represents the in-phase real signal in the second polarization direction in the third real signal of the transform domain, Q' y (w) represents the orthogonal real signal in the second polarization direction in the third real signal in the transform domain.

[0313] Implementation 2.2:

[0314] This corresponds to the implementation method 1.2 in step S1020 above. The processing flow includes:

[0315] The first real number signals in the P-path transform domain are respectively subjected to clock recovery to obtain second real number signals in the P-path transform domain.

[0316] S1040: Obtain a time-domain complex signal.

[0317] Perform at least polarization compensation processing and INTT processing on the second real number signal in the P-channel transform domain to obtain m-channel complex signal X in the time domain and m-channel complex signal Y in the time domain. There are two possible implementation methods:

[0318] Implementation 3.1: The second real signal in the P-path transform domain is sequentially subjected to polarization compensation processing, INTT processing, and merging processing to obtain m-path complex signals X and m-path complex signals Y in the time domain. As shown in Figure 10(c), Figure 10(c) is a schematic flow chart of obtaining a time domain complex signal provided by an embodiment of the present application. The following steps are included: S1041 to S1043:

[0319] S1041 , obtaining a sixth real number signal in a P-way transform domain.

[0320] Performing polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain;

[0321] S1042: Obtain a seventh real number signal in the P-channel time domain.

[0322] Performing INTT processing on the sixth real-number signals in the P-channel transform domain to obtain the seventh real-number signals in the P-channel time domain;

[0323] S1043 , combining to obtain m-path complex signals X and m-path complex signals Y.

[0324] Merging every two of the 2*m-path seventh real-number signals in the time domain in the first polarization direction in the P-path seventh real-number signals to obtain an m-path complex-number signal X in the time domain in the first polarization direction;

[0325] Every two of the 2*m-path seventh real-number signals in the time domain in the second polarization direction in the P-path seventh real-number signals in the time domain are combined to obtain m-path complex signals Y in the time domain in the second polarization direction.

[0326] Implementation 3.2: The second real signal in the P-path transform domain is sequentially subjected to INTT processing, merging processing, and polarization compensation processing. As shown in FIG10( d ), FIG10( d ) is a schematic flow chart of another method for obtaining a time-domain complex signal provided by an embodiment of the present application. The following steps S1044 to S1046 are included:

[0327] S1044 , obtain an eighth real-number signal in the P-channel time domain.

[0328] The second real number signal in the P-path transform domain is subjected to INTT processing to obtain an eighth real number signal in the P-path time domain;

[0329] S1045 , obtaining m complex signals X and m complex signals Y.

[0330] Merging every two of the 2*m eighth real-number signals in the time domain in the first polarization direction in the P eighth real-number signals in the time domain to obtain an m complex-number signal X in the time domain in the first polarization direction;

[0331] Combining every two of the 2*m eighth real-number signals in the time domain in the second polarization direction in the P eighth real-number signals in the time domain to obtain m complex-number signals Y in the time domain in the second polarization direction;

[0332] S1046 , obtaining m-channel complex signals X and m-channel complex signals Y after polarization compensation.

[0333] Polarization compensation is performed on the m-path complex signal X in the time domain in the first polarization direction and the m-path complex signal Y in the time domain in the second polarization direction to obtain the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain after polarization compensation.

[0334] S1050: Obtain a bit signal.

[0335] In order to obtain a recovered bit signal, it is necessary to perform phase recovery and decoding on the time domain complex signal obtained in the above S1040.

[0336] First, phase recovery is performed on the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain to obtain the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain after phase recovery;

[0337] Then, the m-channel complex signals X in the time domain and the m-channel complex signals Y in the time domain after phase recovery are decoded to obtain the m-channel complex signals X in the time domain and the m-channel complex signals Y in the time domain after decoding.

[0338] It should be noted that the embodiments of the present application do not limit how to perform phase recovery and decoding on the time domain complex signal. Reference can be made to the process of performing phase recovery and decoding in the process of electrical signal processing by the current polarization multiplexing coherent receiver.

[0339] The embodiment of the present application also provides a coherent receiver, comprising a polarization beam splitter, a mixer, a photodetector and an analog-to-digital converter, and the Rx DSP shown in FIG9(a), FIG9(b), FIG9(c) or FIG9(d). Figure 1 The above is similar to that shown in , so I will not repeat it here.

[0340] The present application also provides a chip in an embodiment, such as Figure 11 As shown, Figure 11 1 is a schematic diagram of a chip 1100 provided in this application. Chip 1100 includes a processor 1110, a memory 1120, and a communication interface 1130. Processor 1110 is coupled to memory 1120. Memory 1120 is configured to store computer programs, instructions, and / or data. Processor 1110 is configured to execute the computer programs, instructions, and / or data stored in memory 1120, thereby implementing the method described in the method embodiments above.

[0341] As a possible implementation, Figure 11 The chip shown may be a signal processing device including a processor 1110 and a communication interface 1130 , wherein the processor 1110 is coupled to the memory via the communication interface 1130 , and the processor 1110 is configured to execute the method in the above method embodiment.

[0342] It should be understood that the above embodiments are merely examples of the signal processing process provided by the present application and do not constitute any limitation on the scope of protection of the present application. Simple variations of other Rx DSP architectures are also within the scope of protection of the present application. The Rx DSP architecture of the present application differs from the existing Rx DSP architecture in that it avoids using the FFT module and IFFT module for signal processing and instead uses the NTT module and the INTT module.

[0343] It should also be understood that in the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments may be consistent and may be referenced to each other, and the technical features in different embodiments may be combined to form new embodiments according to their inherent logical relationships.

[0344] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0345] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

Claims

1. A signal processing method, characterized in that: The method comprises: Get P-way real signal; The P-way real number signal is processed by number theory transformation NTT to obtain a first real number signal in the P-way transformation domain; Performing dispersion compensation on the first real number signal in the P-path transform domain to obtain a third real number signal in the P-path transform domain; Performing clock recovery on the third real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain; performing polarization compensation and inverse number theory transform INTT processing on the second real number signal in the P-path transform domain to obtain an m-path complex signal X in the time domain and an m-path complex signal Y in the time domain; Performing phase recovery and decoding on the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain to obtain a bit signal; Wherein, m and P are positive integers.

2. The method according to claim 1, wherein The P-channel real-number signals include 4*m-channel real-number signals in an m transmission mode, wherein when m is equal to 1, it indicates single-mode transmission, and when m is greater than 1, it indicates multi-mode transmission.

3. The method according to claim 1 or 2, wherein: The P-path real number signal includes an in-phase real number signal I in the first polarization direction. m x and the orthogonal real signal Q m x , the in-phase real number signal I in the second polarization direction m y and the orthogonal real signal Q m y , Performing dispersion compensation on the first real number signal in the P-path transform domain to obtain a third real number signal in the P-path transform domain includes: The dispersion impulse response corresponding to the in-phase real signal I h (t) and the dispersion impulse response Q corresponding to the orthogonal real signal h (t), the dispersion equalization function I corresponding to the transform domain of the in-phase real signal is obtained through NTT processing h (w) and the dispersion equalization function Q in the transform domain corresponding to the orthogonal real signal h (w); Based on the in-phase real number signal I in the first polarization direction in the first real number signal in the P-path transform domain x (w), orthogonal real signal Q x (w), said I h (w) and the Q h (w) determining a third real signal in the 2*m-way transform domain in the first polarization direction in the third real signal in the P-way transform domain; Based on the in-phase real number signal I in the second polarization direction in the first real number signal in the P-path transform domain y (w), orthogonal real signal Q y (w), said I h (w) and the Q h (w) Determine a third real signal in the 2*m-way transform domain in the second polarization direction in the third real signal in the P-way transform domain.

4. The method according to claim 3, wherein The third real-number signal in the transform domain after dispersion compensation and the first real-number signal in the transform domain meet the following requirements: I ’ x (w)=I x (w)·I h (w)-Q x (w)·Q h (w); Q ’ x (w)=Q x (w)·I h (w)+I x (w)·Q h (w); I ’ y (w)=I y (w)·I h (w)-Q y (w)·Q h (w); Q ’ y (w)=Q y (w)·I h (w)+I y (w)·Q h (w); Among them, I x (w) represents the in-phase real number signal in the first polarization direction in the first real number signal of the transform domain, Q x (w) represents the orthogonal real signal in the first polarization direction in the first real signal of the transform domain, I y (w) represents the in-phase real number signal in the second polarization direction in the first real number signal in the transform domain, Q y (w) represents the orthogonal real signal in the second polarization direction in the first real signal in the transform domain, I ’ x (w) represents the in-phase real number signal in the first polarization direction in the third real number signal in the transform domain, Q ’ x (w) represents the orthogonal real signal in the first polarization direction in the third real signal in the transform domain, I ’ y (w) represents the in-phase real number signal in the second polarization direction in the third real number signal in the transform domain, Q ’ y (w) represents the orthogonal real signal in the second polarization direction in the third real signal in the transform domain.

5. The method according to claim 1 or 2, wherein: The performing polarization compensation and inverse number theory transformation INTT processing on the second real number signal in the P-channel transform domain to obtain m-channel complex signals X in the time domain and m-channel complex signals Y in the time domain includes: Performing polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain; Performing INTT processing on the sixth real-number signal in the P-path transform domain to obtain a seventh real-number signal in the P-path time domain; Each two signals of the 2*m-way seventh real number signals in the P-way time domain in the first polarization direction are merged to obtain the m-way complex signal X in the time domain in the first polarization direction, and each two signals of the 2*m-way seventh real number signals in the P-way time domain in the second polarization direction are merged to obtain the m-way complex signal Y in the time domain in the second polarization direction.

6. The method according to claim 5, wherein Performing polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain includes: equalizing and depolarizing the second real number signal of the 2*m-way transform domain in the first polarization direction in the second real number signal of the P-way transform domain to obtain a sixth real number signal of the 2*m-way transform domain in the first polarization direction in the sixth real number signal of the P-way transform domain; The second real number signal of the 2*m-way transform domain in the second polarization direction in the second real number signal of the P-way transform domain is equalized and depolarized to obtain the sixth real number signal of the 2*m-way transform domain in the second polarization direction in the sixth real number signal of the P-way transform domain.

7. The method according to claim 1 or 2, wherein: The performing polarization compensation and inverse number theory transformation INTT processing on the second real number signal in the P-channel transform domain to obtain m-channel complex signals X in the time domain and m-channel complex signals Y in the time domain includes: Performing INTT processing on the second real number signal in the P-path transform domain to obtain an eighth real number signal in the P-path time domain; Merging each two of the 2*m-path eighth real signals in the P-path time domain in the first polarization direction to obtain an m-path complex signal X in the time domain in the first polarization direction, and merging each two of the 2*m-path eighth real signals in the P-path time domain in the second polarization direction to obtain an m-path complex signal Y in the time domain in the second polarization direction; Polarization compensation is performed in the time domain on the m-channel complex signal X in the first polarization direction and the m-channel complex signal Y in the time domain in the second polarization direction to obtain polarization-compensated m-channel complex signal X in the time domain and m-channel complex signal Y in the time domain.

8. A signal processing method, characterized in that: The method comprises: Get P-way real signal; The P-channel real-number signal is subjected to digital back propagation (DBP) processing to obtain a tenth real-number signal in the P-channel time domain; The tenth real number signal in the P-path time domain is subjected to NTT processing to obtain a first real number signal in the P-path transform domain; Performing clock recovery on the first real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain; Performing polarization compensation and inverse number theory transformation INTT processing on the second real number signal in the P-channel transform domain to obtain m-channel complex number signals X in the time domain and m-channel complex number signals Y in the time domain; Performing phase recovery and decoding on the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain to obtain a bit signal; Wherein, m and P are positive integers.

9. The method according to claim 8, wherein The P-channel real-number signals are subjected to DBP processing to obtain a P-channel time-domain tenth real-number signal, which includes: The P-way real number signal is subjected to NTT processing to obtain a fourth real number signal in a P-way transform domain; The fourth real number signal in the P-path transform domain is subjected to dispersion compensation in the transform domain to obtain a fifth real number signal in the P-path transform domain; The fifth real number signal in the P-path transform domain is subjected to INTT processing to obtain a ninth real number signal in the P-path time domain; The ninth real-number signal in the P-channel time domain is subjected to nonlinear compensation to obtain a tenth real-number signal in the P-channel time domain.

10. The method according to claim 8 or 9, characterized in that The P-channel real-number signals include 4*m-channel real-number signals in an m transmission mode, wherein when m is equal to 1, it indicates single-mode transmission, and when m is greater than 1, it indicates multi-mode transmission.

11. The method according to claim 8 or 9, characterized in that The performing polarization compensation and inverse number theory transformation INTT processing on the second real number signal in the P-channel transform domain to obtain m-channel complex signals X in the time domain and m-channel complex signals Y in the time domain includes: Performing polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain; Performing INTT processing on the sixth real-number signal in the P-path transform domain to obtain a seventh real-number signal in the P-path time domain; Each two signals of the 2*m-way seventh real number signals in the P-way time domain in the first polarization direction are merged to obtain the m-way complex signal X in the time domain in the first polarization direction, and each two signals of the 2*m-way seventh real number signals in the P-way time domain in the second polarization direction are merged to obtain the m-way complex signal Y in the time domain in the second polarization direction.

12. The method according to claim 11, wherein Performing polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain includes: equalizing and depolarizing the second real number signal of the 2*m-way transform domain in the first polarization direction in the second real number signal of the P-way transform domain to obtain a sixth real number signal of the 2*m-way transform domain in the first polarization direction in the sixth real number signal of the P-way transform domain; The second real number signal of the 2*m-way transform domain in the second polarization direction in the second real number signal of the P-way transform domain is equalized and depolarized to obtain the sixth real number signal of the 2*m-way transform domain in the second polarization direction in the sixth real number signal of the P-way transform domain.

13. The method according to claim 8 or 9, characterized in that The performing polarization compensation and inverse number theory transformation INTT processing on the second real number signal in the P-channel transform domain to obtain m-channel complex signals X in the time domain and m-channel complex signals Y in the time domain includes: Performing INTT processing on the second real number signal in the P-path transform domain to obtain an eighth real number signal in the P-path time domain; Merging each two of the 2*m-path eighth real signals in the P-path time domain in the first polarization direction to obtain an m-path complex signal X in the time domain in the first polarization direction, and merging each two of the 2*m-path eighth real signals in the P-path time domain in the second polarization direction to obtain an m-path complex signal Y in the time domain in the second polarization direction; Polarization compensation is performed in the time domain on the m-channel complex signal X in the first polarization direction and the m-channel complex signal Y in the time domain in the second polarization direction to obtain polarization-compensated m-channel complex signal X in the time domain and m-channel complex signal Y in the time domain.

14. A signal processing device, characterized in that: include: A first number theory transform (NTT) module is configured to perform NTT processing on the P-channel input signal to obtain a first real number signal in a P-channel transform domain; A first dispersion compensation module, configured to perform dispersion compensation on the first real number signal in the P-path transform domain to obtain a third real number signal in the P-path transform domain; a clock recovery module, configured to perform clock recovery on the third real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain; A polarization compensation module and a first inverse number theory transform INTT module are used to perform polarization compensation and INTT processing on the second real number signal in the P-channel transform domain to obtain an m-channel complex signal X in the time domain and an m-channel complex signal Y in the time domain; A phase recovery module and a decoding module, configured to process the m-channel complex signal X in the time domain and the m-channel complex signal Y in the time domain to obtain a bit signal; Wherein, m and P are positive integers.

15. The device according to claim 14, wherein The P-channel input signals include 4*m-channel real-number signals in an m transmission mode, wherein when m is equal to 1, it indicates single-mode transmission, and when m is greater than 1, it indicates multi-mode transmission.

16. The device according to claim 15, characterized in that The P-channel input signal includes an in-phase real number signal I in the first polarization direction. m x and the orthogonal real signal Q m x and the in-phase real number signal I in the second polarization direction m y and the orthogonal real signal Q m y The first dispersion compensation module is configured to perform dispersion compensation on the first real number signal in the P-path transform domain to obtain a third real number signal in the P-path transform domain, including: The first dispersion compensation module is used to calculate the dispersion equalization function I of the transform domain corresponding to the in-phase real number signal, the orthogonal real number signal, and the in-phase real number signal in the first polarization direction in the first real number signal of the P-path transform domain. h (w) and the dispersion equalization function Q in the transform domain corresponding to the orthogonal real signal h (w) determining a third real signal in the 2*m-way transform domain in the first polarization direction in the third real signal in the P-way transform domain; The first dispersion compensation module is configured to: h (w) and the Q h (w), determining a third real signal in the 2*m-way transform domain in the second polarization direction in the third real signal in the P-way transform domain, Wherein, the I h (w) and the Q h (w) are the dispersion impulse responses I corresponding to the in-phase real signal h (t) and the dispersion impulse response Q corresponding to the orthogonal real signal h (t)NTT processing is obtained.

17. The device according to claim 16, wherein The third real-number signal in the transform domain after dispersion compensation and the first real-number signal in the transform domain meet the following requirements: I ’ x (w)=I x (w)·I h (w)-Q x (w)·Q h (w); Q ’ x (w)=Q x (w)·I h (w)+I x (w)·Q h (w); I ’ y (w)=I y (w)·I h (w)-Q y (w)·Q h (w); Q ’ y (w)=Q y (w)·I h (w)+I y (w)·Q h (w); Among them, I x (w) represents the in-phase real number signal in the first polarization direction in the first real number signal of the transform domain, Q x (w) represents the orthogonal real signal in the first polarization direction in the first real signal of the transform domain, I y (w) represents the in-phase real number signal in the second polarization direction in the first real number signal in the transform domain, Q y (w) represents the orthogonal real signal in the second polarization direction in the first real signal in the transform domain, I ’ x (w) represents the in-phase real number signal in the first polarization direction in the third real number signal in the transform domain, Q ’ x (w) represents the orthogonal real signal in the first polarization direction in the third real signal in the transform domain, I ’ y (w) represents the in-phase real number signal in the second polarization direction in the third real number signal in the transform domain, Q ’ y (w) represents the orthogonal real signal in the second polarization direction in the third real signal in the transform domain.

18. The device according to any one of claims 14 to 17, characterized in that The first dispersion compensation module includes a third NTT module, a first merging module and a multiplication module, wherein the third NTT module is used to perform NTT processing on the dispersion time domain impulse response to obtain a dispersion equalization function and to perform NTT processing on the first real number signal in the P-path transform domain; the multiplication module is used to multiply the first real number signal in the P-path transform domain after the NTT processing by the dispersion equalization function; and the first merging module is used to merge the signals processed by the multiplication module.

19. The device according to any one of claims 14 to 17, characterized in that The device further comprises: Merge modules; The polarization compensation module is configured to perform polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain; The first INTT module is configured to perform INTT processing on the sixth real-number signal in the P-channel transform domain to obtain a seventh real-number signal in the P-channel time domain; The merging module is used to merge every two signals of the 2*m-way seventh real number signals in the time domain in the first polarization direction in the P-way seventh real number signal to obtain the m-way complex number signal X in the time domain in the first polarization direction, and Each two signals of the 2*m-way seventh real number signals in the time domain in the second polarization direction in the P-way seventh real number signals in the time domain are combined to obtain the m-way complex number signals Y in the time domain in the second polarization direction.

20. The device according to claim 19, wherein The polarization compensation module includes a first butterfly filter and a second butterfly filter; The first butterfly filter is configured to equalize and depolarize the second real signal of the 2*m-way transform domain in the first polarization direction in the second real signal of the P-way transform domain, to obtain a sixth real signal of the 2*m-way transform domain in the first polarization direction in the sixth real signal of the P-way transform domain; The second butterfly filter is used to equalize and depolarize the second real number signal of the 2*m-way transform domain in the second polarization direction in the second real number signal of the P-way transform domain to obtain the sixth real number signal of the 2*m-way transform domain in the second polarization direction in the sixth real number signal of the P-way transform domain.

21. The device according to any one of claims 14 to 17, characterized in that The device further comprises: Merge modules; The first inverse number theory transform INTT module is used to perform INTT processing on the second real number signal in the P-path transform domain to obtain an eighth real number signal in the P-path time domain; The merging module is configured to merge each two signals of the 2*m-way eighth real-number signals in the time domain in the first polarization direction in the P-way eighth real-number signals to obtain an m-way complex signal X in the time domain in the first polarization direction, and Merging every two signals of the 2*m-way eighth real-number signals in the time domain in the second polarization direction in the P-way eighth real-number signals in the time domain to obtain m-way complex signals Y in the time domain in the second polarization direction; The polarization compensation module is used to perform polarization compensation on the m-channel complex signal X and the m-channel complex signal Y in the time domain.

22. The device according to claim 21, wherein The polarization compensation module includes a third butterfly filter; The third butterfly filter is used to perform time-domain polarization compensation on the m-path complex signal X in the time domain in the first polarization direction and the m-path complex signal Y in the time domain in the second polarization direction to obtain the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain after polarization compensation.

23. A signal processing device, characterized in that: include: A digital back propagation (DBP) module is used to perform DBP processing on the P-channel real-number signal to obtain a tenth real-number signal in the P-channel time domain; a first number theory transform (NTT) module, configured to perform NTT processing on the tenth real number signal in the P-path time domain to obtain a first real number signal in the P-path transform domain; a clock recovery module, configured to perform clock recovery on the first real number signal in the P-path transform domain to obtain a second real number signal in the P-path transform domain; A polarization compensation module and a first inverse number theory transform INTT module are used to perform polarization compensation and INTT processing on the second real number signal in the P-channel transform domain to obtain an m-channel complex signal X in the time domain and an m-channel complex signal Y in the time domain; A phase recovery module and a decoding module, configured to process the m-channel complex signal X in the time domain and the m-channel complex signal Y in the time domain to obtain a bit signal; Wherein, m and P are positive integers.

24. The device according to claim 23, wherein The P-channel real-number signals include 4*m-channel real-number signals in an m transmission mode, wherein when m is equal to 1, it indicates single-mode transmission, and when m is greater than 1, it indicates multi-mode transmission.

25. The device according to claim 23 or 24, characterized in that The DBP module includes in sequence: a second NTT module, a second dispersion compensation module, a second INTT module, and a nonlinear compensation module; The DBP module, configured to perform DBP processing on the P-channel real-number signals to obtain the P-channel time-domain tenth real-number signals, includes: The second NTT module is configured to perform NTT processing on the P-way real-number signal to obtain a fourth real-number signal in a P-way transform domain; The second dispersion compensation module is configured to perform dispersion compensation on the fourth real signal in the P-path transform domain in the transform domain to obtain a fifth real signal in the P-path transform domain; The second INTT module is used to perform INTT processing on the fifth real number signal in the P-channel transform domain to obtain a ninth real number signal in the P-channel time domain; The nonlinear compensation module is used to perform nonlinear compensation on the ninth real-number signal in the P-channel time domain to obtain a tenth real-number signal in the P-channel time domain.

26. The device according to claim 23 or 24, characterized in that The device further comprises: Merge modules; The polarization compensation module is configured to perform polarization compensation on the second real number signal in the P-path transform domain to obtain a sixth real number signal in the P-path transform domain; The first INTT module is configured to perform INTT processing on the sixth real-number signal in the P-channel transform domain to obtain a seventh real-number signal in the P-channel time domain; The merging module is used to merge every two signals of the 2*m-way seventh real number signals in the time domain in the first polarization direction in the P-way seventh real number signal to obtain the m-way complex number signal X in the time domain in the first polarization direction, and Each two signals of the 2*m-way seventh real number signals in the time domain in the second polarization direction in the P-way seventh real number signals in the time domain are combined to obtain the m-way complex number signals Y in the time domain in the second polarization direction.

27. The device according to claim 26, wherein The polarization compensation module includes a first butterfly filter and a second butterfly filter; The first butterfly filter is configured to equalize and depolarize the second real signal of the 2*m-way transform domain in the first polarization direction in the second real signal of the P-way transform domain, to obtain a sixth real signal of the 2*m-way transform domain in the first polarization direction in the sixth real signal of the P-way transform domain; The second butterfly filter is used to equalize and depolarize the second real number signal of the 2*m-way transform domain in the second polarization direction in the second real number signal of the P-way transform domain to obtain the sixth real number signal of the 2*m-way transform domain in the second polarization direction in the sixth real number signal of the P-way transform domain.

28. The device according to claim 23 or 24, characterized in that The device further comprises: Merge modules; The first inverse number theory transform INTT module is used to perform INTT processing on the second real number signal in the P-path transform domain to obtain an eighth real number signal in the P-path time domain; The merging module is configured to merge each two signals of the 2*m-way eighth real-number signals in the time domain in the first polarization direction in the P-way eighth real-number signals to obtain an m-way complex signal X in the time domain in the first polarization direction, and Merging every two signals of the 2*m-way eighth real-number signals in the time domain in the second polarization direction in the P-way eighth real-number signals in the time domain to obtain m-way complex signals Y in the time domain in the second polarization direction; The polarization compensation module is used to perform polarization compensation on the m-channel complex signal X and the m-channel complex signal Y in the time domain.

29. The device according to claim 28, wherein The polarization compensation module includes a third butterfly filter; The third butterfly filter is used to perform time-domain polarization compensation on the m-path complex signal X in the time domain in the first polarization direction and the m-path complex signal Y in the time domain in the second polarization direction to obtain the m-path complex signal X in the time domain and the m-path complex signal Y in the time domain after polarization compensation.

30. A coherent receiver, characterized in that: include: The device as described in any one of claims 14 to 28, and the polarization beam splitter, mixer, photodetector and analog-to-digital converter, wherein the polarization beam splitter is used to obtain signals in two polarization directions, the mixer is used to mix the signal in the same polarization direction of the two polarization directions output by the polarization beam splitter, the photodetector is used to convert the intensity of the optical signal output by the mixer into the intensity of the electrical signal, and the analog-to-digital converter is used to convert the signal output by the photodetector from an analog signal to a digital signal.

31. A chip, characterized in that: The chip comprises a communication interface, a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to read and execute the computer program stored in the memory, so that the chip implements the method according to any one of claims 1 to 13.

32. A signal processing device, characterized in that include: processor and communication interface; The processor is coupled to the memory via the communication interface, and the processor is configured to execute program code in the memory to implement the method according to any one of claims 1 to 13.

33. A computer-readable storage medium, characterized in that include: The computer readable medium stores a computer program; When the computer program is run on a computer, the computer is caused to execute the method according to any one of claims 1 to 13.

Images