A multiple-input multiple-output equalization selection method, device, equipment and storage medium

Abstract

The application discloses a multiple-input multiple-output equalization selection method and device, equipment and a storage medium. The method comprises the following steps: inserting a known training sequence into a polarization branch corresponding to each spatial mode of a few-mode fiber transmission system; estimating a frequency domain multiple-input multiple-output channel transmission matrix from each sending branch to each receiving branch based on the known training sequence; extracting the coupling strength between different mode groups according to the block structure of the frequency domain multiple-input multiple-output channel transmission matrix; and selecting a hierarchical multiple-input multiple-output equalization architecture or a complete joint multiple-input multiple-output equalization architecture according to the comparison result of the coupling strength and a preset threshold. The application can reduce the digital signal processing operation complexity and hardware resource consumption in a weak coupling scene by using the hierarchical multiple-input multiple-output equalization architecture, and ensure the equalization performance in a strong coupling scene by using the complete joint multiple-input multiple-output equalization architecture.

Description

Technical Field

[0001] This application relates to the field of optical communication technology, and in particular to a method, apparatus, device and storage medium for multiple input multiple output equalization selection. Background Technology

[0002] With the continuous growth of data communication services, mode division multiplexing (MDM) technology based on few-mode fiber (FMF) utilizes different spatial modes and their polarization branches to transmit data in parallel, becoming an important way to improve the transmission capacity of a single fiber. However, in actual transmission, factors such as fiber bending, mechanical disturbances, temperature changes, and imperfections in mode multiplexing or demultiplexing devices can disrupt the orthogonality between different mode groups, leading to dynamic changes in the coupling strength between mode groups. If a fully joint multiple-input multiple-output (MIMO) equalization architecture is used, it will cause unnecessary digital signal processing (DSP) overhead and hardware resource consumption in weakly coupled scenarios. If a hierarchical MIMO equalization architecture is used, it may not be able to fully compensate for the coupling components between mode groups in strongly coupled scenarios, resulting in a decrease in equalization performance.

[0003] In summary, how to solve the problem of unnecessary digital signal processing overhead and hardware resource consumption, which easily leads to a decline in equalization performance, is an urgent problem that needs to be solved by those skilled in the art. Summary of the Invention

[0004] The purpose of this application is to provide a multiple-input multiple-output (MIMO) equalization selection method. This method can reduce the computational complexity and hardware resource consumption of digital signal processing in weakly coupled scenarios by using a hierarchical multiple-input multiple-output equalization architecture, and ensure equalization performance in strongly coupled scenarios by using a complete joint multiple-input multiple-output equalization architecture. Another purpose of this application is to provide a multiple-input multiple-output equalization selection device, equipment, and computer-readable storage medium.

[0005] To solve the above-mentioned technical problems, this application provides the following technical solution: A multi-input multi-output equalization selection method includes: Insert known training sequences into the polarization branches corresponding to the multiple spatial modes of a few-mode fiber transmission system; Based on the known training sequence, estimate the frequency domain multiple input multiple output channel transmission matrix from each transmitting branch to each receiving branch; Based on the block structure of the frequency domain multiple-input multiple-output channel transmission matrix, the coupling strength between different mode groups is extracted; Based on the comparison result between the coupling strength and the preset threshold, a hierarchical multi-input multi-output equalizer architecture or a fully integrated multi-input multi-output equalizer architecture is selected.

[0006] In one specific embodiment of this application, extracting the coupling strength between different mode groups includes: At each frequency point, singular value decomposition is performed on the frequency domain multiple-input multiple-output channel transmission matrix to obtain a left singular matrix, a singular value diagonal matrix, and a right singular matrix; The unitary channel matrix is ​​constructed using the conjugate transpose of the left singular matrix and the right singular matrix; Based on the block structure of the unitary channel matrix, the coupling strength between the different mode groups at the frequency point is calculated.

[0007] In one specific embodiment of this application, extracting the coupling strength between different mode groups includes: At each frequency point, the square of the Frobenius norm of the off-diagonal block matrix corresponding to different mode groups in the frequency domain multiple input multiple output channel transmission matrix is ​​determined as the inter-mode group coupling power at that frequency point. The square of the Frobenius norm of the diagonal block matrix corresponding to each mode group in the frequency domain multiple input multiple output channel transmission matrix is ​​determined as the intra-mode group pass-through power at the frequency point. The ratio of the inter-mode group coupling power to the intra-mode group direct power is determined as the coupling strength at the frequency point.

[0008] In one specific embodiment of this application, based on the comparison result of the coupling strength and a preset threshold, a hierarchical multiple-input multiple-output (MIMO) equalizer architecture or a fully joint MIMO equalizer architecture is selected, including: Within a preset frequency band, the coupling strength at each frequency point is averaged to obtain the average coupling strength; When the average coupling strength is lower than the preset threshold, the hierarchical multi-input multi-output equalization architecture is selected. When the average coupling strength is higher than or equal to the preset threshold, the complete joint multi-input multi-output equalization architecture is selected.

[0009] In one specific embodiment of this application, the coupling strength between different mode groups is extracted, and based on the comparison result of the coupling strength and a preset threshold, a hierarchical multi-input multi-output equalizer architecture or a complete joint multi-input multi-output equalizer architecture is selected, including: At each frequency point, the first coupling strength and the second coupling strength are calculated respectively, wherein the first coupling strength is the coupling strength between the different mode groups obtained by the direct power method, and the second coupling strength is the coupling strength between the different mode groups obtained by the singular value decomposition method. The first coupling strength and the second coupling strength are weighted and summed to obtain the joint coupling strength; Based on the comparison result between the joint coupling strength and the preset threshold, the hierarchical multi-input multi-output equalization architecture or the complete joint multi-input multi-output equalization architecture is selected.

[0010] In one specific embodiment of this application, estimating the frequency domain multiple-input multiple-output channel transmission matrix from each transmitting branch to each receiving branch includes: Find the target signal that corresponds to the known training sequence from the received signal; The target signal is cross-correlated with the known training sequence to obtain the channel impulse response from each transmitting branch to each receiving branch; The channel impulse response is subjected to a Fourier transform to obtain the frequency domain multiple-input multiple-output channel transmission matrix.

[0011] In one specific embodiment of this application, it further includes: When the few-mode fiber transmission system contains more than three spatial modes and / or more than two polarization dimensions, the polarization branches corresponding to each spatial mode are divided into at least two mode groups according to the actual number of spatial modes and polarization dimensions; wherein, the number of branches in each mode group is determined by the number of spatial modes and the number of polarization dimensions. Based on the number of branches in each mode group, the number of layers and the dimension of each layer of the hierarchical multi-input multi-output balancing architecture are determined, and the total dimension of the complete joint multi-input multi-output balancing architecture is determined.

[0012] A multiple-input multiple-output equalizer selection device, comprising: The sequence insertion module is used to insert known training sequences into the polarization branches corresponding to multiple spatial modes in a few-mode fiber transmission system. The matrix estimation module is used to estimate the frequency domain multiple input multiple output channel transmission matrix from each transmitting branch to each receiving branch based on the known training sequence. The coupling strength extraction module is used to extract the coupling strength between different mode groups based on the block structure of the frequency domain multiple input multiple output channel transmission matrix. The equalization architecture selection module is used to select either a hierarchical multi-input multi-output equalization architecture or a complete joint multi-input multi-output equalization architecture based on the comparison result between the coupling strength and a preset threshold.

[0013] A multiple-input multiple-output equalizer selection device, comprising: Memory, used to store computer programs; A processor for implementing the steps of the aforementioned multiple-input multiple-output equalization selection method when executing the computer program.

[0014] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned multiple-input multiple-output equalization selection method.

[0015] The multiple-input multiple-output (MIMO) equalization selection method provided in this application includes: inserting known training sequences into the polarization branches corresponding to multiple spatial modes in a few-mode fiber transmission system; estimating the frequency domain MIMO channel transmission matrix from each transmitting branch to each receiving branch based on the known training sequences; extracting the coupling strength between different mode groups according to the block structure of the frequency domain MIMO channel transmission matrix; and selecting a hierarchical MIMO equalization architecture or a complete joint MIMO equalization architecture based on the comparison result of the coupling strength and a preset threshold.

[0016] As can be seen from the above technical solution, by inserting known training sequences into the polarization branches of each spatial mode at the transmitting end, estimating the frequency domain multiple-input multiple-output channel transmission matrix based on the training sequences at the receiving end, extracting the coupling strength between different mode groups from the block structure of the matrix, and then comparing the coupling strength with a preset threshold to select a hierarchical multiple-input multiple-output equalization architecture or a complete joint multiple-input multiple-output equalization architecture, the multiple-input multiple-output equalization architecture is selected according to the channel coupling state obtained in real time at the receiving end. Therefore, it can solve the problem of unnecessary digital signal processing overhead and hardware resource consumption, which can easily lead to a decrease in equalization performance. It achieves the technical effect of reducing the complexity of digital signal processing and hardware resource consumption in weakly coupled scenarios with a hierarchical multiple-input multiple-output equalization architecture, and ensuring equalization performance in strongly coupled scenarios with a complete joint multiple-input multiple-output equalization architecture.

[0017] Accordingly, this application also provides a multiple input multiple output equalization selection device, equipment, and computer-readable storage medium corresponding to the above-mentioned multiple input multiple output equalization selection method, which have the above-mentioned technical effects, and will not be repeated here. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the implementation of a multi-input multi-output equalization selection method in this application. Figure 2 This is a schematic diagram showing the estimation results of two crosstalk estimation schemes under a condition without ASE noise in an embodiment of this application, as the preset crosstalk changes. Figure 3 This is a schematic diagram of an estimation result after adding ASE noise in an embodiment of this application; Figure 4 This is a schematic diagram comparing the bit error rate performance when traversing different mode group crosstalk under an optical signal-to-noise ratio of 21 dB in an embodiment of this application. Figure 5 This is a structural block diagram of a multiple-input multiple-output equalizer selection device according to an embodiment of this application; Figure 6 This is a structural block diagram of a multiple-input multiple-output equalizer selection device according to an embodiment of this application; Figure 7 This is a schematic diagram of the specific structure of a multiple-input multiple-output equalization selection device provided in an embodiment of this application. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0022] See Figure 1 , Figure 1 This is a flowchart illustrating an implementation of a multiple-input multiple-output (MIMO) equalization selection method according to an embodiment of this application. The method may include the following steps: S101: Insert known training sequences into the polarization branches corresponding to the multiple spatial modes of a few-mode fiber transmission system.

[0023] At the transmitting end of a few-mode fiber transmission system, known training sequences are inserted into the polarization branches corresponding to multiple spatial modes. This few-mode fiber transmission system refers to a multi-dimensional optical communication transmission system based on mode division multiplexing (MDM) constructed using few-mode fiber (FMF). A spatial mode refers to a linearly polarized mode transmitted independently in a few-mode fiber with different field distributions, including the fundamental mode and several higher-order modes. Taking a three-mode six-dimensional system as an example, the spatial modes include LP01 mode, LP11a mode, and LP11b mode. Each spatial mode contains two orthogonal polarization branches, therefore the total number of transmitting branches is six.

[0024] A known training sequence refers to a sequence whose specific values ​​are known in advance to both the transmitter and receiver. Training sequences can be ZC (Zadoff-Chu) sequences, Constant Amplitude Zero Autocorrelation (CAZAC) sequences, m-sequences, Gold sequences, Walsh sequences, Björck sequences, pseudo-random noise sequences, or other training sequences with low correlation characteristics. Different transmitting branches use the same type of training sequence with different cyclic shifts, or use different types of mutually orthogonal training sequences, so that the receiver can distinguish signals from different transmitting branches.

[0025] S102: Based on the known training sequence, estimate the frequency domain multiple-input multiple-output channel transmission matrix from each transmitting branch to each receiving branch.

[0026] After inserting a known training sequence at the transmitting end, the receiving end receives the signal transmitted through a few-mode fiber channel. The training sequence portion of the received signal is cross-correlated with the locally stored known training sequence to obtain the equivalent channel impulse response from each transmitting branch to each receiving branch. Then, a Fourier transform is performed on the impulse response of each sub-channel to convert the time-domain channel impulse response to the frequency domain, obtaining the equivalent frequency domain response at each frequency point. Arranging the equivalent frequency domain responses from all transmitting branches to all receiving branches at each frequency point according to the transmitting and receiving branches constitutes the frequency-domain multiple-input multiple-output channel transmission matrix. ; Wherein, the element in the q-th row and p-th column of the matrix This indicates that the p-th transmitting branch to the q-th receiving branch is at the frequency point. The equivalent frequency domain response at that point.

[0027] The purpose of this step is to provide the input matrix for subsequent crosstalk monitoring; all subsequent crosstalk calculations are based on this matrix. Expand.

[0028] It should be noted that the implementation of Fourier transform includes, but is not limited to, fast Fourier transform or discrete Fourier transform, and no specific limitation is made here.

[0029] S103: Extract the coupling strength between different mode groups based on the block structure of the frequency domain multiple-input multiple-output channel transmission matrix.

[0030] In obtaining the frequency domain multiple-input multiple-output channel transmission matrix Next, the polarization branches of each spatial mode are assigned to their corresponding mode groups based on mode affiliation. Taking a three-mode six-dimensional system as an example, the LP... 01 The two polarization branches of the mode are assigned to the first mode group, and LP is... 11a and LP 11b The mode has four polarization branches and is classified into the second mode group.

[0031] ; ; Among them, branches 1 and 2 correspond to LP 01 The mode has two polarization states; branches 3, 4, 5, and 6 correspond to LP. 11 Pattern groups, including LP 11a and LP 11b And its dual polarization states. Based on this, the frequency-domain multi-input multi-output channel matrix can be written as: ; in, LP 01 Internal response of the pattern group LP 11 Response within the pattern group; and They represent LP respectively 11 To LP 01 LP 01 To LP 11 The intergroup coupling response.

[0032] At this point, the frequency domain multiple-input multiple-output channel transmission matrix can be naturally divided into diagonal and off-diagonal block matrices both row-wise and column-wise. Therefore, the mode coupling crosstalk monitored in this embodiment is mainly the power leakage ratio of the two off-diagonal blocks relative to the diagonal through block.

[0033] Based on the above block structure, the coupling strength between different mode groups is extracted. Coupling strength is used to quantify the degree of signal leakage between different mode groups. There are several ways to extract the coupling strength between different mode groups. For example, the coupling strength can be calculated as the ratio of the sum of the power of each element in the off-diagonal block matrix to the sum of the power of each element in the diagonal block matrix. Alternatively, the coupling strength can be calculated as the ratio of the sum of the singular values ​​of the off-diagonal block matrix to the sum of the singular values ​​of the diagonal block matrix. It can also be calculated as the ratio of the Frobenius norm of the off-diagonal block matrix to the Frobenius norm of the diagonal block matrix. Another approach is to first normalize the frequency domain MIMO channel transmission matrix by rows or columns, and then extract the coupling strength based on the normalized matrix using any of the above methods. Normalization methods include, but are not limited to, removing singular value matrices after performing Singular Value Decomposition (SVD) on the matrix, QR decomposition, extreme decomposition, matrix whitening, column normalization, or combined row and column normalization.

[0034] In addition to the Frobenius norm power ratio, other matrix energy indices that characterize the coupling strength between mode groups can be used to calculate coupling strength, such as the sum of matrix element power, normalized power ratio, singular value ratio, eigenvalue index, or cross-correlation coefficient.

[0035] S104: Based on the comparison results between the coupling strength and the preset threshold, select either a hierarchical multi-input multi-output equalizer architecture or a fully integrated multi-input multi-output equalizer architecture.

[0036] After extracting the coupling strength, it is compared with a preset threshold. Based on the comparison result, the subsequent MIMO (Multi-Input Multiple-Output) equalization architecture is selected. The specific value of the preset threshold can be set according to the actual situation. The preset threshold can be determined through offline calibration or online calibration based on the modulation format, target bit error rate or forward error correction threshold, optical signal-to-noise ratio, mode-dependent loss, differential mode group delay, and coupling strength estimation error. When the threshold is set to a lower (more negative) value, the system will switch to the full joint MIMO equalization architecture earlier to increase performance margin.

[0037] When the coupling strength is below a preset threshold, it indicates that the coupling between different mode groups is weak, and the current link can be considered a weakly coupled scenario. In this case, a hierarchical MIMO architecture is selected, meaning that the polarization branches within each mode group undergo independent MIMO equalization processing, without joint equalization of branches across mode groups. For example, in a three-mode six-dimensional system, a 2×2+4×4 hierarchical MIMO architecture, i.e., LP, is used. 01 Within the pattern group, 2×2 multiple-input multiple-output equalization is performed, LP 11The pattern group performs 4×4 multiple-input multiple-output (MIMO) equalization processing. Due to the smaller dimension of the equalization matrix in the hierarchical MIMO equalization architecture, the required complex multiplication and addition operations, equalization tap coefficient storage, and coefficient update complexity are significantly reduced compared to the full joint MIMO equalization architecture, thereby reducing DSP computational overhead and hardware resource consumption.

[0038] The specific form of a hierarchical multi-input multi-output (MIMO) equalization architecture is not limited to 2×2+4×4. Depending on the actual number of spatial modes and the mode grouping method, the hierarchical MIMO equalization architecture can be a three-layer architecture (2×2+2×2+2×2), a two-layer architecture (4×4+4×4), or any combination of mode groups. The equalization algorithms for a complete joint MIMO equalization architecture can employ zero-forcing algorithms, least mean square error algorithms, recursive least squares algorithms, least mean square algorithms, frequency domain equalization algorithms, time domain equalization algorithms, etc.

[0039] When the coupling strength is higher than or equal to a preset threshold, it indicates that the coupling between different mode groups is strong, and the current link belongs to a strongly coupled scenario. In this case, a complete joint MIMO equalization architecture is selected, meaning that global joint MIMO equalization is performed between all transmitting and receiving branches to fully compensate for the off-diagonal coupling components between different mode groups. For example, in a three-mode six-dimensional system, a 6×6 complete joint MIMO equalization architecture is used.

[0040] As can be seen from the above technical solution, by inserting known training sequences into the polarization branches of each spatial mode at the transmitting end, estimating the frequency domain multiple-input multiple-output channel transmission matrix based on the training sequences at the receiving end, extracting the coupling strength between different mode groups from the block structure of the matrix, and then comparing the coupling strength with a preset threshold to select a hierarchical multiple-input multiple-output equalization architecture or a complete joint multiple-input multiple-output equalization architecture, the multiple-input multiple-output equalization architecture is selected according to the channel coupling state obtained in real time at the receiving end. Therefore, it can solve the problem of unnecessary digital signal processing overhead and hardware resource consumption, which can easily lead to a decrease in equalization performance. It achieves the technical effect of reducing the complexity of digital signal processing and hardware resource consumption in weakly coupled scenarios with a hierarchical multiple-input multiple-output equalization architecture, and ensuring equalization performance in strongly coupled scenarios with a complete joint multiple-input multiple-output equalization architecture.

[0041] It should be noted that, based on the above embodiments, this application also provides corresponding improvement solutions. In subsequent embodiments, steps that are the same as or corresponding to those in the above embodiments can be referred to each other, and the corresponding beneficial effects can also be referred to each other. These improvements will not be elaborated upon in the following improved embodiments.

[0042] In one specific embodiment of this application, extracting the coupling strength between different mode groups includes: At each frequency point, singular value decomposition is performed on the frequency domain multiple-input multiple-output channel transmission matrix to obtain the left singular matrix, the singular value diagonal matrix, and the right singular matrix; The unitary channel matrix is ​​constructed using the conjugate transpose of the left and right singular matrices. Based on the block structure of the unitary channel matrix, the coupling strength between different mode groups at a frequency point is calculated.

[0043] Let the frequency domain multiple-input multiple-output channel transmission matrix be a matrix. At each frequency point, singular value decomposition is performed on the matrix to obtain the left singular matrix, the singular value diagonal matrix, and the right singular matrix: ; in, and It is a unitary matrix. It is a singular value diagonal matrix.

[0044] Remove the singular value matrix, retain only the left and right unitary matrices, and construct the normalized channel matrix: ; Since the conjugate transposes of the left and right singular matrices are both unitary matrices, a unitary channel matrix is ​​constructed using their conjugate transposes. Each column vector of the unitary channel matrix has a magnitude of 1, eliminating the influence of gain differences between different channels. This allows the subsequently extracted coupling strength to better reflect the coupling relationship between mode groups in the mode space, rather than the overall effect dominated by the power differences between channels.

[0045] Based on the block structure of the unitary channel matrix, the coupling strength between different mode groups at a given frequency point is calculated using methods such as power ratio, singular value ratio, or Frobenius norm ratio. Taking the Frobenius norm ratio as an example, after dividing the unitary channel matrix into blocks according to mode groups, the ratio of the square of the Frobenius norm of the off-diagonal block matrix to the square of the Frobenius norm of the diagonal block matrix is ​​taken as the coupling strength at that frequency point.

[0046] By first performing singular value decomposition on the frequency domain multi-input multi-output channel transmission matrix, removing the singular value diagonal matrix and retaining only the left and right unitary matrices to construct a unitary channel matrix, and then extracting the coupling strength between different mode groups based on the block structure of the unitary channel matrix, the interference of mode correlation loss and channel gain imbalance on the coupling strength extraction can be weakened. This makes the extracted coupling strength more accurately reflect the spatial coupling relationship between mode groups, thereby improving the robustness and accuracy of coupling strength extraction in scenarios with low optical signal-to-noise ratio or channel power imbalance.

[0047] In one specific embodiment of this application, extracting the coupling strength between different mode groups includes: At each frequency point, the square of the Frobenius norm of the off-diagonal block matrix corresponding to different mode groups in the frequency domain multiple input multiple output channel transmission matrix is ​​determined as the inter-mode group coupling power at the frequency point. The square of the Frobenius norm of the diagonal block matrix corresponding to each mode group in the frequency domain multi-input multi-output channel transmission matrix is ​​determined as the intra-mode group pass-through power at the frequency point. The ratio of the coupling power between mode groups to the direct power within a mode group is determined as the coupling strength at a frequency point.

[0048] For a three-mode six-dimensional system, the frequency-domain multiple-input multiple-output channel transmission matrix is ​​a 6x6 square matrix. The 2×2 submatrix corresponding to rows 1-2 and columns 1-2 is used as the diagonal block matrix for the first mode group. The 4×4 submatrix corresponding to rows 3-6 and columns 3-6 is used as the diagonal block matrix for the second mode group. The 2×4 submatrix corresponding to rows 1-2 and columns 3-6, and the 4×2 submatrix corresponding to rows 3-6 and columns 1-2 are used as the off-diagonal block matrix.

[0049] At each frequency point, the square of the Frobenius norm of the off-diagonal block matrix corresponding to different mode groups in the frequency domain multiple-input multiple-output channel transmission matrix is ​​determined as the inter-mode group coupling power at that frequency point: ; in, This represents the inter-mode group coupling power at that frequency point in the direct power method. This represents a 4×2 off-diagonal block matrix from the LP11 transmit mode group to the LP01 receive mode group. This represents a 2×4 off-diagonal block matrix from the LP01 transmit mode group to the LP11 receive mode group; the first subscript of the block indicates the receive mode group, and the second subscript indicates the transmit mode group.

[0050] The square of the Frobenius norm of the diagonal block matrix corresponding to each mode group in the frequency domain multiple-input multiple-output channel transmission matrix is ​​used to determine the intra-mode group cut-through power at a frequency point: ; in, This indicates the pass-through power within the mode group at that frequency point. LP 01 The diagonal block matrix of the pattern group LP 11 The diagonal block matrix of the pattern group.

[0051] The ratio of inter-mode group coupling power to intra-mode group through power is defined as the coupling strength at a frequency point. ; in, Representing frequency point Linear power ratio.

[0052] Its dB form is: ; To reduce the impact of fluctuations at a single frequency point on the judgment result, the embodiments of this application focus on the effective frequency band. Internal calculation of average crosstalk: ; in, This represents the set of valid frequency points participating in the statistics. Indicates the number of valid frequency points.

[0053] The Frobenius norm is the square root of the sum of the squares of the moduli of all elements in a matrix. Its square is the sum of the squares of the moduli of each element in the matrix, reflecting the total energy or power of the corresponding matrix.

[0054] By directly calculating the inter-mode group coupling power and intra-mode group pass-through power based on the off-diagonal and diagonal block matrices of the frequency domain MIMO channel transmission matrix, and using the ratio of these two values ​​as the coupling strength between different mode groups, no additional matrix decomposition or normalization is required. This method is computationally inefficient, simple to implement, and suitable for scenarios with high signal-to-noise ratios and relatively balanced channel gains. Furthermore, since the direct power method is calculated entirely based on the channel transmission matrix actually estimated at the receiver, the resulting coupling strength directly reflects the actual effective coupling level faced by the receiver.

[0055] In one specific embodiment of this application, a hierarchical multiple-input multiple-output (MIMO) equalizer architecture or a fully joint MIMO equalizer architecture is selected based on a comparison between the coupling strength and a preset threshold, including: Within a preset frequency band, the coupling strength at each frequency point is averaged to obtain the average coupling strength; When the average coupling strength is lower than a preset threshold, a hierarchical multi-input multi-output equalization architecture is selected. When the average coupling strength is higher than or equal to a preset threshold, the full joint multi-input multi-output equalization architecture is selected.

[0056] Since the coupling strength at a single frequency point may fluctuate due to local noise, channel estimation errors, or frequency-selective channel response at that frequency point, making decisions based solely on the coupling strength at a single frequency point may lead to unstable or erroneous handovers in the MIMO architecture. Therefore, an arithmetic average of the coupling strength at each frequency point is performed within a preset frequency band.

[0057] ; in, This represents the frequency-by-frequency mode group coupling power in the SVD method. LP represents the average SVD method 11 Pattern group to LP 01 The off-diagonal block matrix of the pattern group, LP represents the average SVD method 01 Pattern group to LP 11 The off-diagonal block matrix of the pattern group.

[0058] ; in, This represents the pass-through power within the mode group at that frequency point using the average SVD method. LP represents the average SVD method 01 The diagonal block matrix of the pattern group LP represents the average SVD method 11 The diagonal block matrix of the pattern group.

[0059] The ratio of inter-mode group coupling power to intra-mode group through power is defined as the coupling strength at a frequency point. ; in, Representing frequency point The coupling strength at that point.

[0060] .

[0061] in, This represents the average inter-mode group coupling strength obtained by first calculating the arithmetic mean of the linear values ​​of the SVD coupling strength at each frequency point within the effective frequency band, and then converting it to dB. The preset frequency band range refers to the pre-defined effective signal frequency band range, which can be determined based on the system's signal bandwidth and occupied spectrum.

[0062] Then, the average coupling strength is compared with a preset threshold. When the average coupling strength is lower than the preset threshold, the current link is determined to be in a weakly coupled state, and a hierarchical MIMO equalizer architecture is selected. When the average coupling strength is higher than or equal to the preset threshold, the current link is determined to be in a strongly coupled state, and a fully coupled MIMO equalizer architecture is selected.

[0063] It should be noted that the preset threshold can be set and adjusted according to the actual situation, and this application embodiment does not limit this. Regarding the selection method of the threshold, in addition to the aforementioned fixed threshold and threshold adjusted according to the link status, the threshold can also be dynamically adjusted according to the modulation format, transmission distance, target bit error rate (BER), MDL size, training sequence estimation reliability, or system complexity requirements. For example, in actual deployment, offline calibration or online calibration can be performed first to establish a mapping table between OSNR and reasonable thresholds. The receiver can then look up the corresponding threshold in the table based on the currently estimated OSNR. Multiple alternative thresholds can also be set and tried sequentially during the system initialization phase. The threshold that makes the received BER meet the target requirements and has the lowest DSP complexity can be selected as the final threshold used.

[0064] By averaging the coupling strength of each frequency point within a preset frequency band to obtain the average coupling strength, and comparing the average coupling strength rather than the coupling strength of a single frequency point with a preset threshold to select a hierarchical or fully joint multi-input multi-output equalization architecture, the influence of coupling strength fluctuations caused by local noise or channel estimation errors at a single frequency point on the architecture selection decision can be effectively suppressed, reducing the probability of false handover and improving the stability and reliability of multi-input multi-output equalization architecture selection.

[0065] In one specific embodiment of this application, the coupling strength between different mode groups is extracted, and based on the comparison result of the coupling strength and a preset threshold, a hierarchical multi-input multi-output equalizer architecture or a complete joint multi-input multi-output equalizer architecture is selected, including: At each frequency point, the first coupling strength and the second coupling strength are calculated respectively, wherein the first coupling strength is the coupling strength between different mode groups obtained by the direct power method, and the second coupling strength is the coupling strength between different mode groups obtained by the singular value decomposition method. The first coupling strength and the second coupling strength are weighted and summed to obtain the joint coupling strength; Based on the comparison results between the joint coupling strength and the preset threshold, either a hierarchical multi-input multi-output equalizer architecture or a complete joint multi-input multi-output equalizer architecture is selected.

[0066] The direct power method calculates coupling strength directly based on the block structure of the frequency-domain MIMO channel transmission matrix. Its advantages include low computational cost and the ability to reflect the actual end-to-end effective coupling level faced by the receiver. However, its estimation accuracy may decrease in scenarios with low optical signal-to-noise ratio (SNR) or unbalanced channel power. The singular value decomposition (SVD) method weakens the impact of channel gain differences by performing singular value decomposition on the frequency-domain MIMO channel transmission matrix and constructing a unitary channel matrix. It is more robust in scenarios with low SNR or unbalanced channel power, but introduces additional matrix decomposition operations. This application combines the advantages of both methods in its embodiments.

[0067] ; in, This represents the joint coupling strength, where α is the weighting coefficient, and its value ranges from 0 to 1.

[0068] The weighting coefficient α is determined based on the actual link status. When the OSNR is high and the channel gain is relatively balanced, the estimation accuracy of the direct power method is sufficient, and α can be set to a large value, such as 0.8, meaning the joint coupling strength is mainly determined by the result of the direct power method. When the OSNR is low or the channel power imbalance is significant, the robustness advantage of the singular value decomposition method becomes apparent, and α can be set to a smaller value, such as 0.2, meaning the joint coupling strength is mainly determined by the result of the singular value decomposition method. In scenarios where the type of the current link cannot be determined, α can be set to 0.5 by default, i.e., equal weighted average.

[0069] After obtaining the joint coupling strength, it is compared with a preset threshold. When the joint coupling strength is lower than the preset threshold, a hierarchical MIMO equalizer architecture is selected. When the joint coupling strength is higher than or equal to the preset threshold, a full joint MIMO equalizer architecture is selected.

[0070] By independently calculating the coupling strength at each frequency point using the direct power method and the singular value decomposition method, and then weighting and summing the results of the two methods to obtain the joint coupling strength, the advantages of the direct power method (low computational cost and accurate reflection of actual coupling level) and the singular value decomposition method (strong robustness in low OSNR scenarios) can be combined. This overcomes the limitations of a single method under different link conditions, ensuring that the joint coupling strength maintains good accuracy and stability under various link conditions, thereby improving the reliability of the multi-input multi-output equalization architecture selection.

[0071] In one specific embodiment of this application, estimating the frequency domain multiple-input multiple-output channel transmission matrix from each transmitting branch to each receiving branch includes: Find the target signal that corresponds to the known training sequence from the received signal; The target signal is cross-correlated with the known training sequence to obtain the channel impulse response from each transmitting branch to each receiving branch; The channel impulse response is subjected to Fourier transform to obtain the frequency domain multiple-input multiple-output channel transmission matrix.

[0072] After completing frame synchronization, the receiving end locates the start and end positions of the training sequence in the received signal based on the known frame structure, and extracts the target signal corresponding to the known training sequence from the received signal, i.e., the target signal segment. This target signal segment is the received version of the training sequence after transmission through the few-mode fiber channel.

[0073] The target signal of each received branch is cross-correlated with the known training sequences inserted by each branch at the transmitting end. The cross-correlation process is as follows: for the target signal of a certain received branch and the known training sequence of a certain transmitted branch, the known training sequence is time-shifted by sampling points and then multiplied with the target signal to obtain the cross-correlation value at each time delay position. The cross-correlation values ​​at all time delay positions constitute the channel impulse response from the transmitted branch to the received branch. Since the training sequences of different transmitted branches have low cross-correlation, the cross-correlation operation can separate the contributions from different transmitted branches to the channel impulse response of each received branch.

[0074] After obtaining the channel impulse responses from all transmit branches to all receive branches, a Fourier transform is performed on the impulse response sequence of each sub-channel to convert the time-domain impulse response to the frequency domain, obtaining the equivalent frequency domain response of that sub-channel at each frequency point. The frequency domain responses from all transmit branches to all receive branches at all frequency points are then combined to form a frequency-domain multiple-input multiple-output channel transmission matrix. The implementation methods of the Fourier transform include, but are not limited to, Fast Fourier Transform or Discrete Fourier Transform.

[0075] By locating the target signal corresponding to the training sequence from the received signal, and using the cross-correlation operation between the target signal and the known training sequence to obtain the channel impulse response from each transmitting branch to each receiving branch, and then obtaining the frequency domain multi-input multi-output channel transmission matrix through Fourier transform, the low cross-correlation characteristics of the training sequence are fully utilized to perform parallel separation and extraction of the channel response of multiple branches. No additional pilot or probe signal overhead is required, making the implementation simple and the channel estimation accuracy high. This provides an accurate channel matrix basis for subsequent coupling strength extraction and selection of multi-input multi-output equalization architecture.

[0076] In one specific embodiment of this application, it further includes: When a few-mode fiber transmission system contains more than three spatial modes and / or more than two polarization dimensions, the polarization branches corresponding to each spatial mode are divided into at least two mode groups according to the actual number of spatial modes and polarization dimensions; wherein, the number of branches in each mode group is determined by the number of spatial modes and the number of polarization dimensions. Based on the number of branches in each pattern group, determine the number of layers and the dimension of each layer in the hierarchical multi-input multi-output balancing architecture, and determine the total dimension of the complete joint multi-input multi-output balancing architecture.

[0077] The foregoing embodiments have been described primarily using a three-mode, six-dimensional system as an example, but the solutions in this application are also applicable to systems with more spatial modes and more polarization dimensions. When the number of spatial modes is greater than three, for example, including LP... 01 LP 11a LP 11b and LP 21 Four or more spatial modes can be grouped into the same mode group based on the similarity of their propagation constants. When there are more than two polarization dimensions, such as using multiple orthogonal polarization states, the polarization branches of the same spatial mode are grouped into the mode group corresponding to that spatial mode.

[0078] After determining the number of branches for each mode group, the number of layers in the hierarchical MIMO equalization architecture is equal to the number of mode groups, and the dimension of the equalization matrix for each layer is the number of branches for the corresponding mode group. For example, when the system contains four spatial modes, each with two polarization branches, for a total of eight branches, if LP... 01 Divide into two branches in the first mode group, and LP 11a and LP 11b Divided into four branches in the second mode group, LP 21 Divided into two branches in the third mode group, the hierarchical MIMO equalization architecture is a three-layer architecture of 2×2+4×4+2×2. The total dimension of the complete joint MIMO equalization architecture is the total number of transmitting branches multiplied by the total number of receiving branches.

[0079] It should be noted that the above method of dividing pattern groups is only an example. In practical applications, multiple spatial patterns can be merged into one pattern group as needed, or different sub-patterns of a single spatial pattern can be split into multiple pattern groups. No specific limitations are made here, with the principle of being able to flexibly perform multi-input multi-output processing at different levels according to the actual coupling state.

[0080] By linking the mode group division and the determination of the dimensions of the multiple-input multiple-output equalization architecture with the actual number of spatial modes and polarization dimensions of the system, the scheme of this application is not only applicable to three-mode six-dimensional few-mode fiber systems, but can also be extended to multi-dimensional optical communication systems containing more spatial modes and more polarization dimensions. It has good versatility and scalability, and can adapt to the future development trend of few-mode fiber systems expanding to higher-order modes.

[0081] To verify the effectiveness of the proposed solution, the following supplementary simulation results include specific parameters and conclusions. A three-mode, six-dimensional few-mode fiber transmission simulation system was constructed, with the simulation signal using 42 GBaud 16-QAM format. The system includes LP... 01 LP 11a LP 11b Three spatial modes and their dual polarization branches are introduced, along with a differential mode group delay of 200 ps and controllable crosstalk.

[0082] See Figure 2 , Figure 2 This is a schematic diagram showing the estimation results of two crosstalk estimation schemes under a condition without ASE noise loading, as the preset crosstalk changes. Figure 2 The figure shows the estimation results of two coupling strength estimation schemes under conditions without ASE noise, as the preset coupling strength changes. As can be seen from the figure, both estimated value curves follow the theoretical coupling strength value well, with estimation errors close to zero. This indicates that under ideal signal-to-noise ratio conditions, the scheme of this application can accurately extract the coupling strength between different mode groups.

[0083] See Figure 3 , Figure 3 This is a schematic diagram of the estimation result after adding ASE noise in an embodiment of this application. Figure 3 The estimation results after adding ASE noise are shown. Within a strong coupling range (greater than -15 dB), the estimates from both methods still accurately follow the theoretical coupling strength. In a weak coupling range (less than -20 dB), the estimation error of the direct power method increases relatively due to the noise floor and channel estimation error. The singular value decomposition method, by unitary normalizing the channel matrix, can mitigate the impact of channel power imbalance and noise disturbances on coupling strength estimation, exhibiting better stability and estimation accuracy in the weak coupling range.

[0084] See Figure 4 , Figure 4 This is a schematic diagram comparing the bit error rate performance when traversing different mode group crosstalk under an optical signal-to-noise ratio of 21 dB in an embodiment of this application. Figure 4The comparison of BER performance under different mode group coupling strengths at an OSNR of 21 dB is further demonstrated. It can be seen that when the coupling strength is below the decision threshold, the BER of the hierarchical MIMO architecture is basically the same as that of the full joint MIMO architecture. When the coupling strength exceeds the threshold, the BER of the hierarchical MIMO architecture increases significantly, while the full joint MIMO architecture remains good. This indicates that the hierarchical MIMO architecture can meet the performance requirements under weak coupling, but the full joint MIMO architecture must be switched to under strong coupling. Under the simulation conditions of 42 GBaud 16-QAM and an OSNR of 21 dB, some branches reach or exceed the HD-FEC threshold at approximately -27 dB, and the BER approaches or exceeds this threshold at approximately -26 dB. Therefore, the switching threshold is set to -28 dB to reserve redundancy for coupling strength estimation errors, noise fluctuations, and implementation deviations; the hierarchical architecture is used when the estimated coupling strength is below -28 dB, and the full 6×6 MIMO architecture is switched to when it is above or equal to -28 dB.

[0085] Corresponding to the above method embodiments, this application also provides a multiple-input multiple-output equalizer selection device. The multiple-input multiple-output equalizer selection device described below can be referred to in correspondence with the multiple-input multiple-output equalizer selection method described above.

[0086] See Figure 5 , Figure 5 This is a structural block diagram of a multiple-input multiple-output equalizer selection device according to an embodiment of this application. The device may include: Sequence insertion module 51 is used to insert known training sequences into the polarization branches corresponding to the multiple spatial modes of a few-mode fiber transmission system. The matrix estimation module 52 is used to estimate the frequency domain multiple input multiple output channel transmission matrix from each transmitting branch to each receiving branch based on the known training sequence. The coupling strength extraction module 53 is used to extract the coupling strength between different mode groups based on the block structure of the frequency domain multiple input multiple output channel transmission matrix. The equalization architecture selection module 54 is used to select either a hierarchical multi-input multi-output equalization architecture or a fully joint multi-input multi-output equalization architecture based on the comparison result between the coupling strength and the preset threshold.

[0087] As can be seen from the above technical solution, by inserting known training sequences into the polarization branches of each spatial mode at the transmitting end, estimating the frequency domain multiple-input multiple-output channel transmission matrix based on the training sequences at the receiving end, extracting the coupling strength between different mode groups from the block structure of the matrix, and then comparing the coupling strength with a preset threshold to select a hierarchical multiple-input multiple-output equalization architecture or a complete joint multiple-input multiple-output equalization architecture, the multiple-input multiple-output equalization architecture is selected according to the channel coupling state obtained in real time at the receiving end. Therefore, it can solve the problem of unnecessary digital signal processing overhead and hardware resource consumption, which can easily lead to a decrease in equalization performance. It achieves the technical effect of reducing the complexity of digital signal processing and hardware resource consumption in weakly coupled scenarios with a hierarchical multiple-input multiple-output equalization architecture, and ensuring equalization performance in strongly coupled scenarios with a complete joint multiple-input multiple-output equalization architecture.

[0088] In one specific embodiment of this application, the coupling strength extraction module 53 may include: The matrix acquisition submodule is used to perform singular value decomposition on the frequency domain multi-input multi-output channel transmission matrix at each frequency point to obtain the left singular matrix, the singular value diagonal matrix, and the right singular matrix; The matrix construction submodule is used to construct the unitary channel matrix using the conjugate transpose of the left and right singular matrices; The first coupling strength calculation submodule is used to calculate the coupling strength between different mode groups at frequency points based on the block structure of the unitary channel matrix.

[0089] In one specific embodiment of this application, the coupling strength extraction module 53 may include: The coupling power determination submodule is used to determine the inter-mode group coupling power at each frequency point by taking the square of the Frobenius norm of the off-diagonal block matrix corresponding to different mode groups in the frequency domain multi-input multi-output channel transmission matrix. The through power determination submodule is used to determine the square of the Frobenius norm of the diagonal block matrix corresponding to each mode group in the frequency domain multi-input multi-output channel transmission matrix as the through power of the mode group at the frequency point. The coupling strength determination submodule is used to determine the coupling strength at a frequency point by the ratio of the coupling power between mode groups to the direct power within a mode group.

[0090] In one specific embodiment of this application, the balanced architecture selection module 54 may include: The average coupling strength acquisition submodule is used to average the coupling strength at each frequency point within a preset frequency band to obtain the average coupling strength. The first equalization architecture selection submodule is used to select a hierarchical multi-input multi-output equalization architecture when the average coupling strength is lower than a preset threshold. The second equalization architecture selection submodule is used to select a complete joint multi-input multi-output equalization architecture when the average coupling strength is higher than or equal to a preset threshold.

[0091] In one specific embodiment of this application, the coupling strength extraction module 53 may include: The second coupling strength calculation submodule is used to calculate the first coupling strength and the second coupling strength at each frequency point, wherein the first coupling strength is the coupling strength between different mode groups obtained by the direct power method, and the second coupling strength is the coupling strength between different mode groups obtained by the singular value decomposition method. The joint coupling strength acquisition submodule is used to perform a weighted summation of the first coupling strength and the second coupling strength to obtain the joint coupling strength; The balanced architecture selection module 54 is specifically a module that selects a hierarchical multi-input multi-output balanced architecture or a complete joint multi-input multi-output balanced architecture based on the comparison result of the joint coupling strength and the preset threshold.

[0092] In one specific embodiment of this application, the matrix estimation module 52 may include: The signal search submodule is used to search for the target signal that corresponds to the known training sequence from the received signal; The channel impulse response acquisition submodule is used to perform cross-correlation operation between the target signal and the known training sequence to obtain the channel impulse response from each transmit branch to each receive branch; The channel transmission matrix acquisition submodule is used to perform Fourier transform on the channel impulse response to obtain the frequency domain multi-input multi-output channel transmission matrix.

[0093] In one specific embodiment of this application, the device may further include: The mode grouping module is used to divide the polarization branches corresponding to each spatial mode into at least two mode groups when the few-mode fiber transmission system contains more than three spatial modes and / or more than two polarization dimensions, based on the actual number of spatial modes and polarization dimensions; wherein the number of branches in each mode group is determined by the number of spatial modes and the number of polarization dimensions. The total dimension determination module is used to determine the number of layers and the dimension of each layer of the hierarchical multi-input multi-output balanced architecture based on the number of branches of each pattern group, and to determine the total dimension of the complete joint multi-input multi-output balanced architecture.

[0094] For the method embodiments described above, see [link to relevant documentation]. Figure 6 , Figure 6This is a schematic diagram of the multiple-input multiple-output equalizer selection device provided in this application. The device may include: Memory 332 is used to store computer programs; The processor 322 is used to implement the steps of the multiple-input multiple-output equalization selection method in the above method embodiments when executing a computer program.

[0095] For details, please refer to Figure 7 , Figure 7 This is a schematic diagram illustrating the specific structure of a multiple-input multiple-output (MIMO) equalization selection device provided in this embodiment. This MIMO equalization selection device can vary significantly due to different configurations or performance characteristics. It may include a processor (central processing unit, CPU) 322 (e.g., one or more processors) and a memory 332. The memory 332 stores one or more computer programs 342 or data 344. The memory 332 can be temporary or persistent storage. The program stored in the memory 332 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the data processing device. Furthermore, the processor 322 may be configured to communicate with the memory 332 and execute the series of instruction operations stored in the memory 332 on the MIMO equalization selection device 301.

[0096] The multiple input multiple output equalizer selection device 301 may also include one or more power supplies 326, one or more wired or wireless network interfaces 350, one or more input / output interfaces 358, and / or one or more operating systems 341.

[0097] The steps in the multiple-input multiple-output equalization selection method described above can be implemented by the structure of a multiple-input multiple-output equalization selection device.

[0098] Corresponding to the above method embodiments, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the following steps: In a few-mode fiber transmission system, known training sequences are inserted into the polarization branches corresponding to the various spatial modes. Based on the known training sequences, the frequency domain multiple-input multiple-output (MIMO) channel transmission matrix from each transmit branch to each receive branch is estimated. According to the block structure of the frequency domain MIMO channel transmission matrix, the coupling strength between different mode groups is extracted. Based on the comparison between the coupling strength and a preset threshold, a hierarchical MIMO equalization architecture or a complete joint MIMO equalization architecture is selected.

[0099] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0100] For a description of the computer-readable storage medium provided in this application, please refer to the above method embodiments; further details will not be repeated here.

[0101] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatuses, devices, and computer-readable storage media disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0102] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the technical solutions and core ideas of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A multi-input multi-output equalization selection method, characterized in that, include: Insert known training sequences into the polarization branches corresponding to the multiple spatial modes of a few-mode fiber transmission system; Based on the known training sequence, estimate the frequency domain multiple input multiple output channel transmission matrix from each transmitting branch to each receiving branch; Based on the block structure of the frequency domain multiple-input multiple-output channel transmission matrix, the coupling strength between different mode groups is extracted; Based on the comparison result between the coupling strength and the preset threshold, a hierarchical multi-input multi-output equalizer architecture or a fully integrated multi-input multi-output equalizer architecture is selected.

2. The multiple-input multiple-output equalization selection method according to claim 1, characterized in that, Extracting the coupling strength between different mode groups, including: At each frequency point, singular value decomposition is performed on the frequency domain multiple-input multiple-output channel transmission matrix to obtain a left singular matrix, a singular value diagonal matrix, and a right singular matrix; The unitary channel matrix is ​​constructed using the conjugate transpose of the left singular matrix and the right singular matrix; Based on the block structure of the unitary channel matrix, the coupling strength between the different mode groups at the frequency point is calculated.

3. The multiple-input multiple-output equalization selection method according to claim 1, characterized in that, Extracting the coupling strength between different mode groups, including: At each frequency point, the square of the Frobenius norm of the off-diagonal block matrix corresponding to different mode groups in the frequency domain multiple input multiple output channel transmission matrix is ​​determined as the inter-mode group coupling power at that frequency point. The square of the Frobenius norm of the diagonal block matrix corresponding to each mode group in the frequency domain multiple input multiple output channel transmission matrix is ​​determined as the intra-mode group pass-through power at the frequency point. The ratio of the inter-mode group coupling power to the intra-mode group direct power is determined as the coupling strength at the frequency point.

4. The multiple-input multiple-output equalization selection method according to claim 2 or 3, characterized in that, Based on the comparison result between the coupling strength and the preset threshold, a hierarchical multi-input multi-output equalizer architecture or a fully joint multi-input multi-output equalizer architecture is selected, including: Within a preset frequency band, the coupling strength at each frequency point is averaged to obtain the average coupling strength; When the average coupling strength is lower than the preset threshold, the hierarchical multi-input multi-output equalization architecture is selected. When the average coupling strength is higher than or equal to the preset threshold, the complete joint multi-input multi-output equalization architecture is selected.

5. The multiple-input multiple-output equalization selection method according to claim 1, characterized in that, Extracting the coupling strength between different mode groups, and selecting a hierarchical multi-input multi-output equalizer architecture or a fully joint multi-input multi-output equalizer architecture based on the comparison result of the coupling strength with a preset threshold, including: At each frequency point, the first coupling strength and the second coupling strength are calculated respectively, wherein the first coupling strength is the coupling strength between the different mode groups obtained by the direct power method, and the second coupling strength is the coupling strength between the different mode groups obtained by the singular value decomposition method. The first coupling strength and the second coupling strength are weighted and summed to obtain the joint coupling strength; Based on the comparison result between the joint coupling strength and the preset threshold, the hierarchical multi-input multi-output equalization architecture or the complete joint multi-input multi-output equalization architecture is selected.

6. The multiple-input multiple-output equalization selection method according to claim 1, characterized in that, Estimate the frequency-domain multiple-input multiple-output channel transmission matrix from each transmit branch to each receive branch, including: Find the target signal that corresponds to the known training sequence from the received signal; The target signal is cross-correlated with the known training sequence to obtain the channel impulse response from each transmitting branch to each receiving branch; The channel impulse response is subjected to a Fourier transform to obtain the frequency domain multiple-input multiple-output channel transmission matrix.

7. The multiple-input multiple-output equalization selection method according to claim 1, characterized in that, Also includes: When the few-mode fiber transmission system contains more than three spatial modes and / or more than two polarization dimensions, the polarization branches corresponding to each spatial mode are divided into at least two mode groups according to the actual number of spatial modes and polarization dimensions; wherein, the number of branches in each mode group is determined by the number of spatial modes and the number of polarization dimensions. Based on the number of branches in each mode group, the number of layers and the dimension of each layer of the hierarchical multi-input multi-output balancing architecture are determined, and the total dimension of the complete joint multi-input multi-output balancing architecture is determined.

8. A multiple-input multiple-output equalizer selection device, characterized in that, include: The sequence insertion module is used to insert known training sequences into the polarization branches corresponding to multiple spatial modes in a few-mode fiber transmission system. The matrix estimation module is used to estimate the frequency domain multiple input multiple output channel transmission matrix from each transmitting branch to each receiving branch based on the known training sequence. The coupling strength extraction module is used to extract the coupling strength between different mode groups based on the block structure of the frequency domain multiple input multiple output channel transmission matrix. The equalization architecture selection module is used to select either a hierarchical multi-input multi-output equalization architecture or a complete joint multi-input multi-output equalization architecture based on the comparison result between the coupling strength and a preset threshold.

9. A multiple-input multiple-output equalizer selection device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the multiple-input multiple-output equalization selection method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the multiple-input multiple-output equalization selection method as described in any one of claims 1 to 7.

Images