High-Order Polarization Mode Dispersion Monitoring Method and Device Based on Fractional Fourier Transform

Abstract

The present invention provides a method for monitoring high-order polarization mode dispersion based on fractional Fourier transform: inserting a training sequence into the transmitted signal, encoding the transmitted signal at the transmitting end, and using upsampling and pulse shaping to compress the signal bandwidth; transmitting the transmitted signal through a long-distance polarization multiplexing coherent transmission link, and performing polarization hierarchical coherent sampling on the high-speed coherent optical signal transmitted through the system at the receiving end by a coherent receiver; sampling the coherent received signal and intercepting the training sequence, and performing fractional Fourier transform and inverse Radon transform on it to reconstruct the time-frequency evolution of the optical signal into a two-dimensional time-frequency image in the standard rectangular coordinate system; combining the two-dimensional time-frequency image with machine learning to extract high-order polarization damage features, and calculating the high-order polarization mode dispersion value of the optical fiber communication system. The present invention realizes the accurate estimation of high-order polarization mode dispersion. The present invention also provides a corresponding device for monitoring high-order polarization mode dispersion based on fractional Fourier transform.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical fiber communication systems, and more specifically, relates to a method and device for monitoring high-order polarization mode dispersion based on fractional-order Fourier transform. Background Art

[0002] Since the 21st century, coherent optical communication, which is mainly characterized by polarization multiplexing, high-order modulation, coherent detection and digital signal processing, has become the core technology of the next generation of ultra-high-speed, large-capacity, and long-distance optical fiber communication systems. With the rapid development of information services such as the Internet, streaming media, and big data, the number of network users has increased dramatically, and the transmission data traffic will show explosive growth. Optical fiber communication systems are constantly developing in the direction of ultra-high speed, large bandwidth, and high spectrum efficiency. In order to improve efficiency while ensuring the quality of signal transmission, signal damage monitoring and compensation are very critical.

[0003] The transmission performance of current long-distance, high-capacity optical fiber communication networks is mainly limited by polarization mode dispersion damage. When the system capacity increases, high-order polarization mode dispersion causes the system performance to deteriorate. Therefore, it is very important to analyze the characteristics of high-order polarization damage in network links and monitor the damage size. Current high-capacity, long-distance optical communication transmission systems are facing extremely complex channel damage. With the accumulation of large dispersion, strong nonlinearity, amplifier spontaneous emission noise, and inter-channel crosstalk, the existing polarization damage detection and performance evaluation methods cannot resist severe ASE noise and nonlinear effects, and it is difficult to meet system requirements. It is urgent to study the time-frequency law of high-order polarization mode dispersion damage from the perspective of joint processing in the time-frequency domain. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention provides a high-order polarization mode dispersion monitoring method based on fractional Fourier transform, which aims to perform damage analysis and accurate estimation of high-order polarization dispersion from a two-dimensional perspective in the time-frequency domain, thereby solving the technical problem that optical performance monitoring tends to fail under complex channel damage.

[0005] To achieve the above object, according to one aspect of the present invention, a method for monitoring high-order polarization mode dispersion based on fractional-order Fourier transform is provided, comprising the following steps:

[0006] (1) Insert a training sequence into the transmitted signal, encode the transmitted signal, and use upsampling and pulse shaping to compress the signal bandwidth;

[0007] (2) The transmitting end signal obtained in step (1) is transmitted using a long-distance polarization multiplexing coherent transmission link, and the high-speed coherent optical signal transmitted through the system is subjected to polarization-graded coherent sampling at the receiving end by a coherent receiver;

[0008] (3) Sample the coherent received signal obtained in step (2) and intercept the training sequence, and perform fractional Fourier transform and inverse Radon transform on it to reconstruct the time-frequency evolution of the optical signal into a two-dimensional time-frequency image in the standard rectangular coordinate system;

[0009] (4) Combine the two-dimensional time-frequency image obtained in step (3) with machine learning to extract high-order polarization damage features, and calculate the high-order polarization mode dispersion value of the optical fiber communication system.

[0010] In one embodiment of the present invention, the step (1) includes:

[0011] (1.1) Insert a training sequence into the transmitter signal to construct a time-frequency distribution image;

[0012] (1.2) Encode the transmitter signal inserted with the training sequence, and the signal is selected as PDM-16QAM;

[0013] (1.3) Adopt twice upsampling and RRC shaping with a shaping factor of 0.1 to compress the transmitter signal bandwidth.

[0014] In one embodiment of the present invention, the step (2) includes:

[0015] (2.1) Drive the IQ modulator with the transmitter signal obtained in step (1) to perform signal modulation and load it onto the optical carrier to obtain the transmitter optical signal;

[0016] (2.2) Transmit the transmitter optical signal obtained in step (2.1) through a long-distance polarization multiplexing coherent transmission link;

[0017] (2.3) Perform polarization hierarchical coherent sampling on the transmitted optical signal to obtain a coherent received signal.

[0018] In one embodiment of the present invention, the step (3) includes:

[0019] (3.1) Downsample the coherent received signal to obtain a digital signal and intercept the training sequence;

[0020] (3.2) Perform fractional Fourier transform scans of different orders on the training sequence of the coherent reception to obtain the time-frequency distribution image of the optical communication signal in polar coordinates;

[0021] (3.3) Perform inverse Radon transform on the time-frequency distribution image in polar coordinates to reconstruct the time-frequency evolution of the optical signal into a two-dimensional time-frequency image in the standard rectangular coordinate system.

[0022] In one embodiment of the present invention, the step (4) includes:

[0023] (4.1)Based on the two-dimensional time-frequency image and combined with the long-haul polarization multiplexing coherent transmission link, establish its corresponding high-order polarization mode dispersion value label;

[0024] (4.2)Read the two-dimensional time-frequency image and the high-order polarization mode dispersion label data obtained in step (4.1) as the input and label value of the neural network, and train the neural network;

[0025] (4.3)Call the trained neural network to extract the high-order polarization damage characteristics and obtain the high-order polarization mode dispersion value of the optical fiber communication system.

[0026] In one embodiment of the present invention, in the step (1.1): perform a 0.3-order fractional Fourier transform on the DC signal to obtain a training sequence. The transformed signal will become a signal similar to chirp, which appears as a straight line with linearly changing frequency over time in the time-frequency domain plane.

[0027] In one embodiment of the present invention, in the step (2.2): in terms of the optical fiber link, build an optical fiber loop to simulate long-distance transmission. The loop module is used to control the number of times passing through the loop. In each loop, there is a 100-km standard single-mode optical fiber and an erbium-doped optical fiber amplifier respectively. After exiting the loop, a multi-segment short optical fiber cascade model is used to simulate the situation of high-order polarization mode dispersion, and the optical signal-to-noise ratio module is used to control the optical signal-to-noise ratio. The signal is affected by nonlinearity, additive Gaussian white noise, and polarization damage during transmission, so as to study the performance of the high-order polarization mode dispersion estimation algorithm under complex channel damage conditions in the ultra-long-haul high baud rate system.

[0028] In one embodiment of the present invention, in the step (3.2): perform fractional Fourier transform scans of different orders on the training sequence. The scan order range is from -1 to 1, and the scan order interval is set to 0.05. The data obtained after scanning different orders are stored in the matrix row by row.

[0029] In one embodiment of the present invention, in the step (3.3): take the square absolute value of the obtained matrix and substitute it into the inverse Radon function. Use linear interpolation to perform inverse Radon transform on the data from -90° to obtain a signal standard time-frequency distribution image with a size of 100×100×2.

[0030] According to another aspect of the present invention, there is also provided a high-order polarization mode dispersion monitoring device based on fractional Fourier transform, including at least one processor and a memory. The at least one processor and the memory are connected through a data bus. The memory stores instructions executable by the at least one processor. After being executed by the processor, the instructions are used to complete the high-order polarization mode dispersion monitoring method based on fractional Fourier transform.

[0031] Generally speaking, compared with the prior art, the above technical solution conceived by the present invention has the following beneficial effects:

[0032] (1) The present invention realizes the accurate estimation of high-order polarization mode dispersion, which helps to improve the performance and reliability of large-capacity systems to meet the growing communication needs;

[0033] (2) The present invention uses time-frequency domain reconstruction technology to visualize the transmission evolution law of optical communication signals in the form of a two-dimensional image, more intuitively and comprehensively analyze the time-frequency distribution law of high-order polarization mode dispersion, accurately extract damage characteristics, and achieve high-precision damage estimation;

[0034] (3) The present invention can realize the accurate estimation of high-order polarization mode dispersion under complex channel impairments, is robust to other channel impairments such as nonlinear impairments and spontaneous emission noise, and is applicable to future complex network architectures. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a schematic diagram of the principle of the high-order polarization mode dispersion monitoring method based on the fractional Fourier transform provided by the present invention;

[0036] Figure 2 is a schematic diagram of the structure of the high-order polarization mode dispersion monitoring method based on the fractional Fourier transform provided by the present invention;

[0037] Figure 3 is a schematic diagram of the process for constructing a time-frequency two-dimensional distribution image based on the fractional Fourier transform provided by the present invention;

[0038] Figure 4 is a time-frequency two-dimensional distribution image based on the fractional Fourier transform provided by the present invention;

[0039] Figure 5 is a schematic diagram of the neural network structure provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0041] As Figure 1 shown, the present invention provides a high-order polarization mode dispersion monitoring method based on the fractional Fourier transform, including:

[0042] (1)Insert a training sequence into the transmitted signal, encode the transmitted signal at the transmitting end, and use upsampling and pulse shaping to compress the signal bandwidth, including:

[0043] (1.1)Insert a training sequence into the transmitted signal at the transmitting end to construct a time-frequency distribution image;

[0044] Specifically, perform a fractional Fourier transform on the DC signal to obtain the training sequence. For a DC signal, its frequency is a fixed value and does not change with time. After performing a fractional Fourier transform on such a signal, due to the time-frequency rotation characteristic, the transformed signal will become a signal similar to a chirp. The transformed signal will become a signal similar to a chirp, which appears as a straight line with frequency linearly changing with time in the time-frequency domain plane.

[0045] In this step, the order of the fractional Fourier transform is 0.3, and the length of the training sequence is 100.

[0046] (1.2)Encode the transmitted signal at the transmitting end into which the training sequence has been inserted;

[0047] In this step, the signal symbols adopt the form of 16-quadrature amplitude-phase modulation, and the length of the signal is 2 16 。

[0048] (1.3)Use upsampling and root-raised cosine shaping to compress the bandwidth of the transmitted signal at the transmitting end.

[0049] Specifically, perform two-fold upsampling, and use a root-raised cosine filter with a roll-off factor of 0.1 to shape the signal.

[0050] (2)Transmit the transmitted signal obtained in step (1) through a long-haul polarization-division multiplexing coherent transmission link, and perform polarization-division coherent sampling on the high-speed coherent optical signal transmitted through the system at the receiving end by a coherent receiver;

[0051] (2.1)Drive the IQ modulator with the transmitted signal obtained in step (1) to perform signal modulation and load it onto the optical carrier to obtain the transmitted optical signal;

[0052] Specifically, the transmitted signal includes four inputs of XI, XQ, YI, and YQ into 2 I / Q modulators. The optical signal is split into X and Y polarization states by a polarization beam splitter (PBS), and then the electrical signal is modulated into an orthogonal optical signal through a Mach-Zehnder modulator. Finally, the polarization beam combiner (PBC) combines these two polarization states together to achieve polarization-division multiplexing.

[0053] (2.2)Transmit the transmitted optical signal obtained in step (2.1) through a long-haul polarization-division multiplexing coherent transmission link;

[0054] Specifically, asFigure 2 As shown in the figure, in terms of the optical fiber link, an optical fiber loop is built to simulate long-distance transmission. The loop module is used to control the number of times passing through the loop. In each loop, there is a section of standard single-mode fiber (SMF) with a length of 100 km and an erbium-doped fiber amplifier (EDFA). After exiting the loop, a multi-section short fiber cascading model is used to simulate the situation of high-order polarization mode dispersion, and the optical signal-to-noise ratio module is used to control the optical signal-to-noise ratio. The signal is affected by nonlinearity, additive Gaussian white noise, and polarization damage during transmission, so as to study the performance of the high-order polarization mode dispersion estimation algorithm under complex channel damage conditions in ultra-long-distance high baud rate systems.

[0055] Among them, a 50-section short fiber cascading model is used to simulate the situation of high-order polarization mode dispersion. The variation range of the first-order polarization mode dispersion is 0 - 120 ps, and the variation range of the second-order polarization dispersion is 0 - 3500 ps 2 ; The system baud rate is 75 GBaud, and the sampling rate is 150 GHz; The output optical power of the transmitter varies in the range of 0 - 6 dBm, with a variation interval of 1 dBm; The number of times of the optical fiber loop varies in the range of 1 - 30, that is, the optical fiber transmission distance range is 100 km - 3000 km, the optical fiber dispersion coefficient is 16.89 ps / nm / km, and the optical fiber dispersion variation range is 1689 ps / nm - 50670 ps / nm;

[0056] (2.3) Perform polarization-graded coherent sampling on the transmitted optical signal to obtain a coherent received signal.

[0057] Specifically, the local oscillator light and the received optical signal at the receiving end enter the coherent receiver together. After being processed by demodulation and a photodetector, the output is four electrical signals, namely XI, XQ, YI, and YQ.

[0058] (3) Sample and intercept the training sequence of the coherent received signal obtained in step (2), and perform fractional Fourier transform and inverse Radon transform on it to reconstruct the time-frequency evolution of the optical signal into a time-frequency two-dimensional image in the standard rectangular coordinate system;

[0059] (3.1) Downsample the coherent received signal to obtain a digital signal and intercept the training sequence;

[0060] Specifically, after twice downsampling, intercept the training sequence with a length of the first 100.

[0061] (3.2) Perform fractional Fourier transform scans of different orders on the training sequence of the coherent reception to obtain the time-frequency distribution image of the optical communication signal in polar coordinates;

[0062] Specifically, as Figure 3As shown, perform fractional Fourier transform scans of different orders on the training sequence. The scanning order range is from -1 to 1, and the scanning order interval is set to 0.05. Store the data obtained after scanning different orders into a matrix row by row.

[0063] (3.3) Perform an inverse Radon transform on the time-frequency distribution image in polar coordinates to reconstruct the time-frequency evolution of the optical signal into a two-dimensional time-frequency image in the standard rectangular coordinate system.

[0064] Specifically, as Figure 3 shown, take the square absolute value of the obtained matrix and substitute it into the inverse Radon function. Use linear interpolation to perform an inverse Radon transform on the data from -90°. Obtain a signal standard time-frequency distribution image with a size of 100×100×2, where 100 represents the length and width of the image, which is consistent with the length of the training sequence, and 2 represents the time-frequency images of the X and Y polarization states. The time-frequency images corresponding to various damages are as Figure 4 shown, showing corresponding image features in the two-dimensional time-frequency plane.

[0065] (4) Combine the two-dimensional time-frequency image obtained in step (4) with machine learning to extract high-order polarization damage features, and calculate the high-order polarization mode dispersion value of the optical fiber communication system.

[0066] (4.1) According to the two-dimensional time-frequency image, combined with the long-distance polarization multiplexing coherent transmission link, establish its corresponding high-order polarization mode dispersion value label;

[0067] Specifically, according to the optical communication transmission system in step (2.2), a total of 100,000 groups of data of optical communication signals under complex channel damages are obtained, and construct a data set with their corresponding damage labels. It is planned to divide the data set into 3 categories, namely: 80% for the training set (a total of 80,000 groups of data), 10% for the validation set (a total of 10,000 groups of data), and 10% for the test set (a total of 10,000 groups of data).

[0068] (4.2) Read the two-dimensional time-frequency image and high-order polarization mode dispersion label data obtained in step (4.1) as the input and label values of the neural network, and train the neural network;

[0069] Specifically, as Figure 5As shown, a Transformer neural network is constructed to analyze the time-frequency characteristics of high-order polarization mode dispersion. First, the image is segmented into a series of non-overlapping image patches, with 200 patches of size 10×10. Then, they are projected into a linear embedding layer. To preserve spatial information, positional encoding is added to each image patch, and then the combined input is fed into the Transformer encoder. The Transformer encoder has two sub-layers. The first is the multi-head self-attention mechanism, and the second is a simple position-aware fully-connected feed-forward neural network. Residual connections are adopted around each sub-layer, followed by layer normalization. The overall algorithm is optimized using the Adam optimizer, with a batch size of 256 and 30,000 iterations. Finally, the time-frequency domain characteristics of high-order polarization mode dispersion are extracted from the two-dimensional time-frequency image, and high-precision damage estimation is achieved.

[0070] (4.3) Call the trained neural network to extract high-order polarization damage features and obtain the high-order polarization mode dispersion value of the optical fiber communication system;

[0071] Specifically, the performance of the neural network is verified using the data in the test set to check its generalization ability.

[0072] Further, the present invention also provides a high-order polarization mode dispersion monitoring device based on the fractional Fourier transform, including at least one processor and a memory. The at least one processor and the memory are connected via a data bus. The memory stores instructions executable by the at least one processor. After being executed by the processor, the instructions are used to complete the high-order polarization mode dispersion monitoring method based on the fractional Fourier transform.

[0073] It is easy for those skilled in the art to understand that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A high-order polarization mode dispersion monitoring method based on fractional Fourier transform, characterized in that, The steps are as follows: (1) Insert a training sequence into the transmitted signal, encode the signal at the transmitting end, and use upsampling and pulse shaping to compress the signal bandwidth; (2) Transmit the transmitted signal obtained in step (1) through a long-haul polarization multiplexed coherent transmission link, and perform polarization-graded coherent sampling on the high-speed coherent optical signal transmitted through the system at the receiving end by a coherent receiver; (3) Sample the coherent received signal obtained in step (2) and intercept the training sequence, and perform fractional Fourier transform and inverse Radon transform on it to reconstruct the time-frequency evolution of the optical signal into a two-dimensional time-frequency image in the standard rectangular coordinate system; (4) Combine the two-dimensional time-frequency image obtained in step (3) with machine learning to extract high-order polarization damage features, and calculate the high-order polarization mode dispersion value of the optical fiber communication system; Step (3) includes: (3.1) Downsample the coherent received signal to obtain a digital signal and intercept the training sequence; (3.2) Perform fractional Fourier transform scans of different orders on the training sequence of the coherent reception to obtain the time-frequency distribution image of the optical communication signal in polar coordinates; (3.3) Perform inverse Radon transform on the time-frequency distribution image in polar coordinates to reconstruct the time-frequency evolution of the optical signal into a two-dimensional time-frequency image in the standard rectangular coordinate system; In step (3.2): Perform fractional Fourier transform scans of different orders on the training sequence, the scanning order range is from -1 to 1, the scanning order interval is set to 0.05, and store the data obtained after scanning of different orders into a matrix row by row; In step (3.3): Take the square absolute value of the obtained matrix, substitute it into the inverse Radon function, and perform inverse Radon transform on the data from -90° using linear interpolation to obtain a signal standard time-frequency distribution image with a size of 100×100×2.

2. The high-order polarization mode dispersion monitoring method based on fractional Fourier transform according to claim 1, characterized in that, Step (1) includes: (1.1) Insert a training sequence into the transmitted signal at the transmitting end to construct a time-frequency distribution image; (1.2) Encode the transmitted signal with the training sequence inserted, and the signal is selected as PDM-16QAM; (1.3) Use twice upsampling and RRC shaping with a shaping factor of 0.1 to compress the transmitted signal bandwidth.

3. The high-order polarization mode dispersion monitoring method based on fractional Fourier transform according to claim 2, characterized in that, Step (2) includes: (2.1) Drive the transmitted signal obtained in step (1) by an IQ modulator to perform signal modulation and load it onto an optical carrier to obtain a transmitted optical signal; (2.2) Transmit the transmitted optical signal obtained in step (2.1) through a long-haul polarization multiplexed coherent transmission link; (2.3) Perform polarization-graded coherent sampling on the transmitted optical signal to obtain a coherent received signal.

4. The high-order polarization mode dispersion monitoring method based on fractional Fourier transform according to claim 2, characterized in that, Step (4) includes: (4.1) According to the two-dimensional time-frequency image, combined with the long-haul polarization multiplexed coherent transmission link, establish its corresponding high-order polarization mode dispersion value label; (4.2) Read the two-dimensional time-frequency image and high-order polarization mode dispersion label data obtained in step (4.1) as the input and label values of the neural network, and train the neural network; (4.3) Call the trained neural network to extract high-order polarization damage features to obtain the high-order polarization mode dispersion value of the optical fiber communication system.

5. The high-order polarization mode dispersion monitoring method based on fractional Fourier transform according to claim 2, characterized in that, In step (1.1): Perform a 0.3-order fractional Fourier transform on the DC signal to obtain a training sequence. The transformed signal will become a signal similar to a chirp, which appears as a straight line with linearly changing frequency over time in the time-frequency domain plane.

6. The high-order polarization mode dispersion monitoring method based on fractional Fourier transform according to claim 3, characterized in that, In step (2.2): In terms of the optical fiber link, an optical fiber loop is built to simulate long-distance transmission. The loop module is used to control the number of times passing through the loop. In each loop, there is a 100-km standard single-mode optical fiber and an erbium-doped optical fiber amplifier respectively. After exiting the loop, a multi-segment short optical fiber cascade model is used to simulate the situation of high-order polarization mode dispersion, and the optical signal-to-noise ratio module is used to control the optical signal-to-noise ratio. The signal is affected by nonlinearity, additive Gaussian white noise, and polarization damage during transmission, so as to study the performance of the high-order polarization mode dispersion estimation algorithm under complex channel damage conditions in an ultra-long-distance high baud rate system.

7. A high-order polarization mode dispersion monitoring device based on fractional Fourier transform, characterized in that: It includes at least one processor and a memory, and the at least one processor and the memory are connected through a data bus. The memory stores instructions executable by the at least one processor. After being executed by the processor, the instructions are used to complete the high-order polarization mode dispersion monitoring method based on fractional Fourier transform described in any one of claims 1-6.

Images