A data recovery method and system for power distribution communication network based on optical fiber network

Abstract

The present invention relates to the field of communication technology, and in particular to a data recovery method and system for a power distribution communication network based on an optical fiber network, including extracting features from the signal waveform data at the receiving end to obtain signal waveform distortion features; taking minimization of inherent inter-symbol interference as the optimization goal, performing global optimization on dispersion equalization parameters according to the signal waveform distortion features to obtain dispersion equalization pre-compensation parameters; in an optical fiber communication link, using an adaptive filtering method based on blind equalization to process dynamic inter-symbol interference signals generated by the mismatch of dispersion equalization pre-compensation parameters to obtain an equalized recovery signal; performing parameter estimation on the equalized recovery signal to obtain an original symbol estimation sequence; obtaining an optical fiber channel parameter estimation value; and optimizing the original symbol estimation sequence using the optical fiber channel parameter estimation value to recover the power distribution communication recovery signal. The present invention improves the accuracy of data recovery by combining technologies such as an adaptive filtering method based on blind equalization.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a method and system for recovering data in a power distribution communication network based on an optical fiber network. Background Art

[0002] With the rapid development of information technology, the integration of distribution communication networks and fiber optic networks has become an essential component of modern communication systems. This integration aims to leverage the high-speed transmission capabilities of fiber optic networks and the extensive coverage of distribution communication networks to jointly build an efficient and reliable communication infrastructure. However, in practical applications, the differences in the characteristics of the two networks have brought unprecedented challenges to data transmission. Fiber optic networks, with their low loss and high bandwidth, dominate long-distance transmission. However, the dispersion effect of fiber links has become a key factor affecting signal quality. Dispersion causes optical pulses to gradually broaden during transmission, resulting in waveform distortion and inter-symbol interference, which affects accurate signal reception and interpretation. At the same time, distribution communication networks face the challenge of multipath effects. This is because when signals propagate on distribution lines, they produce multiple reflections and scattering due to path differences, further exacerbating signal distortion. When these two networks are integrated, the dispersion effect of fiber links and the multipath effect of distribution lines superimpose on each other, exacerbating the problem of waveform distortion during signal transmission. This distortion not only leads to inaccurate information transmission but also significantly increases the delay and bit error rate of the communication system, seriously affecting the reliability and effectiveness of the system.

[0003] To address this issue, dispersion compensation modules and equalizers are widely used in signal conditioning. Dispersion compensation modules are designed to offset the dispersion effects in optical fiber links through specific optical designs, while equalizers are used to adjust the frequency response of signals in power distribution and communication networks to reduce distortion caused by multipath. However, in actual engineering implementation, due to the complexity of the two network characteristics and differences in equipment parameters, parameter mismatches between dispersion compensation modules and equalizers often occur. This mismatch not only fails to effectively eliminate signal distortion, but also introduces new inter-symbol interference (ISI), becoming a bottleneck that restricts transmission performance. Therefore, how to effectively address the combined effects of dispersion and multipath effects in the context of the integration of optical fiber networks and power distribution and communication networks, and resolve the ISI problem caused by parameter mismatches between dispersion compensation modules and equalizers, has become a technical challenge that urgently needs to be addressed. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a data recovery method and system for a power distribution communication network based on an optical fiber network.

[0005] In a first aspect, the present invention provides a method for recovering data in a power distribution communication network based on an optical fiber network, the method comprising the following steps:

[0006] Acquiring receiving end signal waveform data on the optical fiber communication link, and performing feature extraction on the receiving end signal waveform data to obtain signal waveform distortion features;

[0007] Taking minimizing the inherent inter-symbol interference caused by channel characteristics as the optimization goal, the dispersion equalization parameters are globally optimized according to the signal waveform distortion characteristics to obtain the dispersion equalization pre-compensation parameters;

[0008] The dispersion equalization pre-compensation parameters are used to compensate for the dispersion effect in the optical fiber communication link, and in the optical fiber communication link, a dynamic inter-symbol interference signal generated by the mismatch of the dispersion equalization pre-compensation parameters is blindly equalized using an adaptive filtering method based on blind equalization to obtain an equalized recovery signal;

[0009] Performing parameter estimation on the equalized restored signal using a maximum likelihood estimation algorithm to obtain an original symbol estimation sequence;

[0010] A two-way channel transmission model based on optical fiber dispersion and multipath effects is established, and optical fiber channel parameter estimation values ​​are obtained through learning the two-way channel transmission model;

[0011] The optical fiber channel parameter estimation value is used to optimize the original symbol estimation sequence to restore the power distribution communication recovery signal.

[0012] In a further embodiment, the step of performing global optimization on the dispersion equalization parameters according to the signal waveform distortion characteristics with minimizing the inherent intersymbol interference caused by the channel characteristics as the optimization goal to obtain the dispersion equalization pre-compensation parameters includes:

[0013] The support vector machine algorithm is used to learn the signal waveform distortion characteristics and predict the fiber dispersion coefficient and power line multipath delay spread data of the current communication link;

[0014] The current communication link distortion degree is quantified based on the fiber dispersion coefficient and power line multipath delay spread data;

[0015] When the current communication link distortion exceeds a preset distortion threshold, the dispersion equalization parameters are globally optimized using a genetic algorithm with the goal of minimizing the inherent intersymbol interference caused by channel characteristics to obtain the dispersion equalization pre-compensation parameters.

[0016] In a further embodiment, the step of globally optimizing the dispersion equalization parameters using a genetic algorithm comprises:

[0017] Generate a set of random dispersion equalization parameter combinations as an initial population; the dispersion equalization parameters include dispersion compensation depth and equalizer tap coefficients;

[0018] The optimization goal is to minimize the inherent inter-symbol interference caused by channel characteristics, and the dispersion equalization parameters are used as the search space of the genetic algorithm;

[0019] In each iteration, the cost function value corresponding to each dispersion equalization parameter combination is calculated to obtain the fitness;

[0020] Crossover and mutation operations are performed according to the fitness, and multiple rounds of iterative updates are performed until the preset convergence conditions are met to obtain the optimal dispersion equalization parameter combination, which is used as the dispersion equalization pre-compensation parameters; the dispersion equalization pre-compensation parameters include the dispersion compensation depth value and the equalizer tap coefficient value.

[0021] In a further embodiment, the step of performing blind equalization processing on a dynamic inter-symbol interference signal generated by a mismatch in dispersion equalization pre-compensation parameters using an adaptive filtering method based on blind equalization to obtain an equalized recovery signal includes:

[0022] receiving a dynamic inter-symbol interference signal generated by a mismatch in dispersion equalization pre-compensation parameters, and calculating a second-order moment and a fourth-order moment of the dynamic inter-symbol interference signal to obtain a statistical characteristic of the interference signal;

[0023] Calculate the autocorrelation function of the interference signal based on its statistical characteristics, and estimate the signal delay characteristics based on the peak position of the autocorrelation function;

[0024] Setting the tap coefficients required by the blind equalizer according to the signal delay characteristics, and iteratively updating the tap coefficients using a blind equalization algorithm;

[0025] In each iterative update, the minimum distance between the signal point on the constellation diagram of the current blind equalizer output signal and the decision domain boundary is calculated to obtain the distortion;

[0026] If the fluctuation of the distortion degree for multiple consecutive periods is less than a preset fluctuation threshold and the distortion degree is lower than a preset first distortion threshold, it is determined that the blind equalizer has reached a stable state, the tap coefficients of the current blind equalizer are locked, and the current equalization state is maintained;

[0027] After the tap coefficients of the current blind equalizer converge and are locked, the dynamic inter-symbol interference signal is filtered using the current blind equalizer according to the locked tap coefficients, and an equalized recovery signal is output.

[0028] In a further embodiment, if the distortion exceeds a preset second distortion threshold or the constellation diagram deflection angle exceeds a preset deflection angle threshold, the lock state of the current blind equalizer tap coefficient is released, and the tap coefficient is re-iteratively updated.

[0029] In a further embodiment, the update step size of the current blind equalizer tap coefficients is adjusted according to the distortion level.

[0030] In a further embodiment, the step of obtaining the estimated value of the optical fiber channel parameter by learning the two-way channel transmission model includes:

[0031] According to the two-way channel transmission model, a two-way channel transmission process is simulated using a split-step Fourier method to obtain channel state data in the optical fiber communication link;

[0032] Extracting optical fiber channel characteristic parameters based on channel status data in an optical fiber communication link; optical fiber channel characteristic parameters include optical fiber dispersion coefficient, group delay, multipath amplitude and multipath delay;

[0033] A recurrent neural network is used to perform nonlinear mapping on the optical fiber channel characteristic parameters to obtain optical fiber channel parameter estimation values.

[0034] In a further embodiment, the step of optimizing the original symbol estimation sequence using the fiber channel parameter estimation value to recover the power distribution communication recovery signal includes:

[0035] Introducing the fiber channel parameter estimation value, decoding the original symbol estimation sequence using the Viterbi decoding algorithm to obtain a symbol decision sequence;

[0036] Encoding the symbol decision sequence to obtain a coded data sequence, and segmenting the coded data sequence to obtain error correction data blocks;

[0037] Obtaining an average burst error length and frequency of the channel according to the estimated values ​​of the optical fiber channel parameters, and determining an interleaving depth according to the average burst error length and frequency of the channel;

[0038] constructing an interleaving matrix according to the interleaving depth, and rearranging the error correction data block using the interleaving matrix to form an interleaved data block;

[0039] The interleaved data block is decoded using a decoding algorithm based on an error correction code to recover a power distribution communication recovery signal.

[0040] In a further embodiment, the step of introducing the fiber channel parameter estimation value and decoding the original symbol estimation sequence using the Viterbi decoding algorithm to obtain the symbol decision sequence includes:

[0041] Constructing a state trellis diagram according to the constraint length and coding structure of the convolutional code encoder at the transmitting end; each state in the state trellis diagram represents a coding state of the convolutional code encoder;

[0042] During decoding, the Hamming distance is weighted by the fiber channel parameter estimation value according to the original symbol estimation sequence, and the path metric for each state in the state trellis diagram to transfer to the next state is calculated using the weighted Hamming distance;

[0043] By backtracking the state lattice graph, finding a path with the minimum cumulative path metric from the initial state to the final state, and obtaining a surviving path;

[0044] The symbol decisions corresponding to each coding state on the surviving path are arranged in sequence to form a symbol decision sequence.

[0045] In a second aspect, the present invention provides a data recovery system for a power distribution communication network based on an optical fiber network, the system comprising:

[0046] A feature extraction module is used to obtain the receiving end signal waveform data on the optical fiber communication link, and perform feature extraction on the receiving end signal waveform data to obtain signal waveform distortion features;

[0047] a compensation analysis module configured to globally optimize dispersion equalization parameters based on the signal waveform distortion characteristics with the goal of minimizing inherent intersymbol interference caused by channel characteristics, thereby obtaining dispersion equalization pre-compensation parameters;

[0048] a compensation optimization module, configured to compensate for the dispersion effect in the optical fiber communication link using the dispersion equalization pre-compensation parameters, and to perform blind equalization processing on a dynamic inter-symbol interference signal generated by a mismatch in the dispersion equalization pre-compensation parameters in the optical fiber communication link using an adaptive filtering method based on blind equalization, to obtain an equalized recovery signal;

[0049] A parameter estimation module, configured to perform parameter estimation on the equalized restored signal using a maximum likelihood estimation algorithm to obtain an original symbol estimation sequence;

[0050] A channel analysis module is used to establish a two-way channel transmission model based on optical fiber dispersion and multipath effects, and obtain optical fiber channel parameter estimates through learning the two-way channel transmission model;

[0051] The data recovery module is used to optimize the original symbol estimation sequence using the fiber channel parameter estimation value to recover the power distribution communication recovery signal.

[0052] The present invention provides a data recovery method and system for a power distribution communication network based on an optical fiber network. The method obtains signal waveform data of a receiving end on an optical fiber communication link and performs feature extraction on the signal waveform data of the receiving end to obtain signal waveform distortion characteristics; minimizes inherent inter-symbol interference caused by channel characteristics as an optimization goal, performs global optimization on dispersion equalization parameters according to the signal waveform distortion characteristics, and obtains dispersion equalization pre-compensation parameters; uses the dispersion equalization pre-compensation parameters to compensate for the dispersion effect in the optical fiber communication link, and uses an adaptive filtering method based on blind equalization in the optical fiber communication link to perform blind equalization processing on a dynamic inter-symbol interference signal generated by a mismatch of the dispersion equalization pre-compensation parameters to obtain an equalized recovery signal; uses a maximum likelihood estimation algorithm to perform parameter estimation on the equalized recovery signal to obtain an original symbol estimation sequence; establishes a two-way channel transmission model based on optical fiber dispersion and multipath effects, and obtains optical fiber channel parameter estimation values ​​through learning the two-way channel transmission model; uses the optical fiber channel parameter estimation values ​​to optimize the original symbol estimation sequence to recover the power distribution communication recovery signal. Compared with existing technologies, this method can more accurately estimate fiber channel parameters through technologies such as a dual-path channel transmission model based on fiber dispersion and multipath effects, significantly improving the accuracy of data recovery in the distribution communication network and ensuring stable operation of the system and reliable data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a flow chart of a method for recovering data in a power distribution communication network based on an optical fiber network provided by an embodiment of the present invention;

[0054] Figure 2 This is a block diagram of a data recovery system for a power distribution communication network based on an optical fiber network provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The following describes the embodiments of the present invention in detail with reference to the accompanying drawings. The embodiments are provided for illustrative purposes only and are not to be construed as limiting the present invention. The accompanying drawings are provided for reference and illustration only and do not constitute a limitation on the scope of protection of the present invention. Many changes may be made to the present invention without departing from the spirit and scope of the present invention.

[0056] refer to Figure 1 , the embodiment of the present invention provides a data recovery method for a power distribution communication network based on an optical fiber network, such as Figure 1 As shown, the method includes the following steps:

[0057] S1. Acquire receiving-end signal waveform data on an optical fiber communication link, and perform feature extraction on the receiving-end signal waveform data to obtain signal waveform distortion features.

[0058] Specifically, at the receiving end of the optical fiber communication system, this embodiment can use a photodetector to convert the received optical signal into an electrical signal, obtain the receiving end signal waveform data on the optical fiber communication link, and preprocess the collected receiving end signal waveform data. The preprocessing includes steps such as denoising and normalization. The time domain features of the preprocessed receiving end signal waveform data are extracted to obtain the time domain features of the signal waveform distortion. Then, the signal is converted from the time domain to the frequency domain through Fourier transform or wavelet transform to obtain its spectrum information. The frequency domain features of the signal waveform distortion are extracted based on the spectrum information. The time domain features of the signal waveform distortion and the frequency domain features of the signal waveform distortion are integrated to obtain the signal waveform distortion features. These features may include zero crossing wavelength, peak-to-peak amplitude, zero crossing wavelength difference, and wave train area.

[0059] S2. With minimizing the inherent inter-symbol interference caused by the channel characteristics as the optimization goal, the dispersion equalization parameters are globally optimized according to the signal waveform distortion characteristics to obtain the dispersion equalization pre-compensation parameters.

[0060] In this embodiment, the step of performing global optimization of dispersion equalization parameters based on the signal waveform distortion characteristics with minimizing the inherent intersymbol interference caused by channel characteristics as the optimization goal to obtain dispersion equalization pre-compensation parameters includes:

[0061] The support vector machine algorithm is used to learn the signal waveform distortion characteristics and predict the fiber dispersion coefficient and power line multipath delay spread data of the current communication link;

[0062] The current communication link distortion degree is quantified based on the fiber dispersion coefficient and power line multipath delay spread data;

[0063] When the current communication link distortion exceeds a preset distortion threshold, the dispersion equalization parameters are globally optimized using a genetic algorithm with the goal of minimizing the inherent intersymbol interference caused by channel characteristics to obtain the dispersion equalization pre-compensation parameters.

[0064] Specifically, this embodiment uses signal waveform distortion features as input, uses a trained support vector machine algorithm for learning, and predicts fiber dispersion coefficient and power line multipath delay spread data. Based on the predicted fiber dispersion coefficient and power line multipath delay spread data, one or more quantitative indicators are calculated to evaluate the degree of distortion of the communication link. For example, this embodiment may use signal quality indicators (such as signal-to-noise ratio, bit error rate) or channel characteristic indicators (such as dispersion length, multipath delay spread) to quantify the degree of distortion. Then, when the current degree of distortion of the communication link exceeds a preset distortion threshold, a genetic algorithm is used to globally optimize the dispersion equalization parameters with the optimization goal of minimizing the inherent intersymbol interference (ISI) caused by the channel characteristics. In this embodiment, the step of globally optimizing the dispersion equalization parameters using the genetic algorithm includes:

[0065] Generate a set of random dispersion equalization parameter combinations as the initial population to form the search space of the genetic algorithm; the dispersion equalization parameters include dispersion compensation depth and equalizer tap coefficients;

[0066] The optimization goal is to minimize the inherent inter-symbol interference caused by channel characteristics, and the dispersion equalization parameters are used as the search space of the genetic algorithm;

[0067] In each iteration, the cost function value corresponding to each dispersion equalization parameter combination is calculated to obtain the fitness;

[0068] Crossover and mutation operations are performed according to the fitness, and multiple rounds of iterative updates are performed until the preset convergence conditions are met to obtain the optimal dispersion equalization parameter combination, which is used as the dispersion equalization pre-compensation parameters; the dispersion equalization pre-compensation parameters include the dispersion compensation depth value and the equalizer tap coefficient value.

[0069] In the process of using a genetic algorithm to globally optimize the dispersion equalization parameters, this embodiment can design a cost function to measure the magnitude of the inherent inter-symbol interference (ISI) caused by channel characteristics. The cost function should be able to reflect the impact of different dispersion equalization parameter configurations on system performance. For example, the cost function can be based on the signal reconstruction error, bit error rate, or other performance indicators related to ISI as the basis for the genetic algorithm to perform selection operations. The genetic algorithm is then initialized to define the value range and encoding method (such as binary encoding, real number encoding, etc.) of the dispersion equalization parameters, and the population is initialized, that is, a set of random dispersion equalization parameters are generated. The parameter combination is used as the initial solution, and the cost function value corresponding to each dispersion equalization parameter combination is calculated to obtain the fitness of each individual in the population. The individuals with higher fitness are selected as parents, crossover and mutation operations are performed to generate new offspring, and the population is updated. The offspring replace some or all of the parents and enter the next round of iteration. The above steps are repeated for multiple rounds of iterative updates to check whether the maximum number of iterations is reached. If the maximum number of iterations is reached, the optimal dispersion equalization parameter combination is output as the dispersion equalization pre-compensation parameter and applied as the dispersion equalization pre-compensation parameter in the communication system, thereby reducing ISI and improving communication quality.

[0070] S3. Use the dispersion equalization pre-compensation parameters to compensate for the dispersion effect in the optical fiber communication link, and use an adaptive filtering method based on blind equalization in the optical fiber communication link to blindly equalize the dynamic inter-symbol interference signal generated by the mismatch of the dispersion equalization pre-compensation parameters to obtain an equalized recovery signal.

[0071] In this embodiment, the step of using a blind equalization-based adaptive filtering method to blindly equalize a dynamic inter-symbol interference signal generated by a dispersion equalization pre-compensation parameter mismatch to obtain an equalized recovery signal includes:

[0072] receiving a dynamic inter-symbol interference signal generated by a mismatch in dispersion equalization pre-compensation parameters, and calculating a second-order moment and a fourth-order moment of the dynamic inter-symbol interference signal to obtain a statistical characteristic of the interference signal;

[0073] Calculate the autocorrelation function of the interference signal based on its statistical characteristics, and estimate the signal delay characteristics based on the peak position of the autocorrelation function;

[0074] Setting the tap coefficients required by the blind equalizer according to the signal delay characteristics, and iteratively updating the tap coefficients using a blind equalization algorithm;

[0075] In each iterative update, the minimum distance between the signal point on the constellation diagram of the current blind equalizer output signal and the decision domain boundary is calculated to obtain the distortion;

[0076] If the fluctuation of the distortion degree for multiple consecutive periods is less than a preset fluctuation threshold and the distortion degree is lower than a preset first distortion threshold, it is determined that the blind equalizer has reached a stable state, the tap coefficients of the current blind equalizer are locked, and the current balanced state is maintained; if the distortion degree exceeds a preset second distortion threshold or the deflection angle of the constellation diagram exceeds a preset deflection angle threshold, the locking state of the tap coefficients of the current blind equalizer is released, and the tap coefficients are iteratively updated again;

[0077] After the tap coefficients of the current blind equalizer converge and are locked, the dynamic inter-symbol interference signal is filtered using the current blind equalizer according to the locked tap coefficients, and an equalized recovery signal is output.

[0078] It should be noted that the dispersion compensation module is used to offset the dispersion effect in the optical fiber link, and the equalizer is used to adjust the frequency response of the signal in the power distribution communication network to reduce the distortion caused by the multipath effect. The differences in the two major network characteristics and equipment parameters lead to parameter mismatch between the dispersion compensation module and the equalizer. This mismatch will introduce new inter-symbol interference, which becomes a bottleneck restricting transmission performance. This embodiment uses a blind equalization algorithm to process the dynamic inter-symbol interference generated by parameter mismatch. In the specific implementation process, the signal delay is first determined by statistical characteristics and the equalizer is initialized; then the blind equalization algorithm is used to iteratively update the tap coefficients of the equalizer, and the convergence status of the equalizer is monitored by distortion and constellation diagram deflection angle; finally, after the equalizer converges, the equalized recovery signal is output. This mechanism can dynamically adapt to channel changes and ensure stable signal quality. The specific implementation process is as follows:

[0079] Dynamic inter-symbol interference signals are usually manifested as the expansion of signals in the time domain due to the mismatch between dispersion and pre-compensation parameters during transmission, resulting in mutual interference between symbols. In this embodiment, the dynamic inter-symbol interference signal generated by the mismatch between dispersion equalization and pre-compensation parameters can be received by a signal receiving device, and the received dynamic inter-symbol interference signal is digitized and converted into a dynamic inter-symbol interference digital signal for subsequent processing. The dynamic inter-symbol interference digital signal is statistically calculated to calculate its second-order moment (i.e., the variance of the signal) and fourth-order moment (i.e., the kurtosis of the signal). The second-order moment reflects the energy distribution of the signal, and the fourth-order moment reflects the sharpness of the signal. The calculated second-order moment and fourth-order moment are used as the statistical characteristics of the interference signal, and its autocorrelation function is calculated based on the statistical characteristics of the interference signal. The signal delay characteristics are estimated by the peak position of the autocorrelation function. The delay characteristic refers to the delay time caused by the dispersion and pre-compensation mismatch during the transmission of the signal. Due to the delay caused by the dispersion and pre-compensation mismatch, the peak position will deviate from the zero delay point. The signal delay characteristics are estimated based on the offset between the peak position and the zero delay point. The tap coefficients required for the blind equalizer (i.e., the order of the filter) are set according to the signal delay characteristics, and the tap coefficients are iteratively updated using a blind equalization algorithm to achieve blind equalization processing of dynamic inter-symbol interference signals.

[0080] In this embodiment, the number of taps should be set to be greater than or equal to the delay value to ensure that all inter-symbol interference can be covered. The tap coefficients of the equalizer are initialized, and the center tap of the equalizer is assigned a unit gain (for example, 1), and the remaining taps are assigned a zero gain. In this way, the initial equalizer is equivalent to a unit gain filter. A blind equalization algorithm is used to iteratively update the tap coefficients of the equalizer. The blind equalization algorithm includes the constant modulus algorithm (CMA) and the multi-mode algorithm (MMA). In each iteration, the minimum distance from the signal point to the decision domain boundary on the constellation diagram of the current equalizer output signal is calculated as a measure of the distortion. The update step size (dynamic step size) of the current blind equalizer tap coefficient is adjusted according to the distortion value. The larger the distortion, the larger the step size to accelerate the convergence speed; the smaller the distortion, The smaller the step size, the better the convergence accuracy. If the distortion fluctuation is less than the preset fluctuation threshold for multiple consecutive cycles and the value is lower than the first distortion threshold, it means that the equalizer has converged, the current tap coefficient can be locked, and further updates can be stopped; if the distortion exceeds the second distortion threshold, or the constellation diagram deflection angle exceeds the preset deflection angle threshold, it means that the channel conditions have changed (for example, the dispersion equalization pre-compensation parameters are further mismatched), and the equalizer tap coefficient needs to be unlocked and adaptive adjustment needs to be restarted. When the equalizer tap coefficient converges and is locked, the current equalizer is used to filter the dynamic inter-symbol interference signal and output an equalized recovery signal. The equalized recovery signal has low inter-symbol interference, and the constellation diagram should be close to the constellation diagram of the ideal signal.

[0081] S4. Perform parameter estimation on the equalized restored signal using a maximum likelihood estimation algorithm to obtain an original symbol estimation sequence.

[0082] Maximum Likelihood Estimation (MLE) is a statistically based parameter estimation method that aims to find parameter values ​​that maximize the probability of observed data. In digital communications, this embodiment can use MLE to estimate the parameters of the equalized restored signal to obtain an estimated sequence of original symbols. Specifically, this embodiment defines a likelihood function, which is the probability of different symbol sequences occurring given the observed data (equalized restored signal). This probability is calculated based on the noise distribution (such as Gaussian noise) in the model and the characteristics of the symbol sequence after passing through the channel. A predetermined search algorithm is then used to search for the symbol sequence that maximizes the likelihood function within a given symbol set. For example, an exhaustive search algorithm (for short sequences) or a Viterbi algorithm (for long sequences) is used to search for the symbol sequence that maximizes the likelihood function. The algorithm undergoes multiple iterations, each of which updates the likelihood value based on the current estimated symbol sequence and the observed data. When the search algorithm converges or reaches a predetermined number of iterations, the symbol sequence with the maximum likelihood value is output as the estimated original symbol sequence.

[0083] S5. A two-way channel transmission model based on optical fiber dispersion and multipath effects is established, and optical fiber channel parameter estimation values ​​are obtained through learning the two-way channel transmission model.

[0084] In this embodiment, the step of obtaining the fiber channel parameter estimation value by learning the two-way channel transmission model includes:

[0085] According to the two-way channel transmission model, a two-way channel transmission process is simulated using a split-step Fourier method to obtain channel state data in the optical fiber communication link;

[0086] Extracting optical fiber channel characteristic parameters based on channel status data in an optical fiber communication link; optical fiber channel characteristic parameters include optical fiber dispersion coefficient, group delay, multipath amplitude and multipath delay;

[0087] A recurrent neural network is used to perform nonlinear mapping on the optical fiber channel characteristic parameters to obtain optical fiber channel parameter estimation values.

[0088] Specifically, this embodiment constructs a two-way channel transmission model based on fiber dispersion and multipath effects. The model should be able to simulate the signal transmission process in the fiber optic communication link, including the influence of dispersion effect and multipath effect on the signal. Fiber dispersion can be simulated by considering the difference in transmission speed of optical signals of different wavelengths in the optical fiber, and the multipath effect can be simulated by introducing multiple paths, each path having different amplitudes and delays. After establishing the two-way channel transmission model, this embodiment adopts the split-step Fourier method to simulate the signal transmission process. Specifically, the split-step Fourier method is used to simulate the propagation of light in the optical fiber, and the initial state of the input signal, including its frequency, amplitude and phase, is set. The split-step Fourier transform is used to perform Fourier transform on each small step of the signal transmitted in the optical fiber, and then the influence of dispersion and multipath effects is considered in the frequency domain. The signal is converted from the frequency domain back to the time domain through the inverse Fourier transform to obtain the signal state in the next step. The above steps are repeated until the signal completes the transmission of the entire optical fiber link, and the channel state data in the optical fiber communication link is obtained. , including information such as the amplitude, phase and delay of the signal. After obtaining the channel state data, this embodiment extracts characteristic parameters of the optical fiber channel from the channel state data. These parameters include optical fiber dispersion coefficient, group delay, multipath amplitude and multipath delay. Among them, the optical fiber dispersion coefficient can be obtained by analyzing the change of the signal in the frequency domain, the group delay can be obtained by measuring the transmission time of the signal at different frequencies, the multipath amplitude can be obtained by analyzing the multipath components of the signal in the time domain, and the multipath delay can be obtained by measuring the arrival time difference of the signals on different paths. Finally, a recurrent neural network (RNN) is used to perform nonlinear mapping on the extracted optical fiber channel characteristic parameters to obtain estimated values ​​of the optical fiber channel parameters. Through the above steps, this embodiment can use a two-way channel transmission model and a recurrent neural network to estimate the parameters of the optical fiber channel. This method can accurately simulate the signal transmission process in the optical fiber communication link and extract key optical fiber channel characteristic parameters, and then perform nonlinear mapping through the neural network to obtain accurate estimated values ​​of the optical fiber channel parameters.

[0089] S6. Optimize the original symbol estimation sequence using the fiber channel parameter estimation value to restore the power distribution communication recovery signal.

[0090] In this embodiment, the step of optimizing the original symbol estimation sequence using the fiber channel parameter estimation value to recover the power distribution communication recovery signal includes:

[0091] Introducing the fiber channel parameter estimation value, decoding the original symbol estimation sequence using the Viterbi decoding algorithm to obtain a symbol decision sequence;

[0092] Encoding the symbol decision sequence to obtain a coded data sequence, and segmenting the coded data sequence to obtain error correction data blocks;

[0093] Obtaining an average burst error length and frequency of the channel according to the estimated values ​​of the optical fiber channel parameters, and determining an interleaving depth according to the average burst error length and frequency of the channel;

[0094] constructing an interleaving matrix according to the interleaving depth, and rearranging the error correction data block using the interleaving matrix to form an interleaved data block;

[0095] The interleaved data block is decoded using a decoding algorithm based on an error correction code to recover a power distribution communication recovery signal.

[0096] In the power distribution communication system, after the signal is transmitted through the optical fiber channel, due to the influence of channel characteristics (such as loss, dispersion, etc.), the original symbol estimation sequence obtained by the receiving end will have certain errors. In order to accurately restore the original signal, it is necessary to optimize the original symbol estimation sequence using the optical fiber channel parameter estimation value, and combine the Viterbi decoding algorithm and interleaving technology to improve the decoding performance and error correction capability. Specifically, this embodiment first uses the Viterbi decoding algorithm to decode the original symbol estimation sequence, and can find the transmission sequence that best matches the received sequence under the given convolution code constraints. After obtaining the symbol decision sequence, it is used to The symbol decision sequence is encoded using the same coding method as the transmitter (such as convolutional code, block code, etc.) to generate a coded data sequence. The coded data sequence is then segmented into fixed-length segments to obtain fixed-length error-corrected data blocks. The channel burst error statistics are calculated based on the fiber channel parameter estimates. The channel burst error statistics include the average burst error length and frequency of the channel. Based on the channel burst error statistics, the interleaving depth of the interleaving matrix is ​​set. The interleaving depth is an important parameter of the interleaving matrix, which determines the degree of data rearrangement during the interleaving process to ensure that the interleaved data blocks can span multiple burst error regions.

[0097] This embodiment rearranges the error correction data blocks by the interleaving depth of the interleaving matrix to form interleaved data blocks. During the interleaving process, it is determined whether the interleaving interval meets the requirements (i.e., the interleaving interval is greater than the average burst error length of the channel) to span the average burst error length of the channel. If the interleaving interval meets the requirements, the current interleaving depth is locked and no longer changed. Based on the interleaved data blocks, a decoding algorithm based on error correction codes (such as the BCJR algorithm, the SOVA algorithm, etc.) is used to decode the interleaved data blocks. During the decoding process, the characteristics of the error correction code are used to detect and correct errors in the data. Finally, the data sequence obtained by decoding is the distribution communication recovery signal.

[0098] In this embodiment, the Viterbi decoding algorithm and the fiber channel parameter estimation value are used to decode the received original symbol estimation sequence to obtain the symbol decision sequence. This process takes into account the impact of the fiber channel on the signal, and evaluates the advantages and disadvantages of different paths through weighted Hamming distance and path metric. Ultimately, the most likely surviving path is found and the symbol decision sequence is extracted. In this embodiment, the steps of introducing the fiber channel parameter estimation value and decoding the original symbol estimation sequence using the Viterbi decoding algorithm to obtain the symbol decision sequence include:

[0099] Constructing a state trellis diagram according to the constraint length and coding structure of the convolutional code encoder at the transmitting end; each state in the state trellis diagram represents the coding state of the convolutional code encoder at a certain moment;

[0100] During decoding, the Hamming distance is weighted by the fiber channel parameter estimation value according to the original symbol estimation sequence, and the path metric for each state in the state trellis diagram to transfer to the next state is calculated using the weighted Hamming distance;

[0101] By backtracking the state lattice graph, finding a path with the minimum cumulative path metric from the initial state to the final state, and obtaining a surviving path;

[0102] The symbol decisions corresponding to each coding state on the surviving path are arranged in sequence to form a symbol decision sequence.

[0103] Specifically, the Viterbi decoding algorithm can correctly traverse the state trellis diagram and recover a more accurate power distribution communication recovery signal by analyzing the constraint length and coding structure of the convolutional code encoder. Assuming that the constraint length of the convolutional code encoder used by the transmitter is 3 and the coding rate is 1 / 2 (that is, for every one bit input, two bits are output), the generator polynomial of the convolutional code encoder can be G1 = 133 (octal) and G2 = 171 (octal). Since the constraint length is 3, the number of states of the convolutional code encoder is = 4 states, which can be marked as S0, S1, S2, and S3. According to the generator polynomial of the convolutional code encoder, the possible output bits and the path to the next state are determined for each state. For example, from the S0 state, according to the input bit 0 or 1, it may be transferred to S0 or S2, and the corresponding bit sequence is output. The state transition diagram is unfolded on the time axis to form a state trellis diagram. Each time step corresponds to an input bit and an output bit sequence. This embodiment introduces the fiber channel parameter estimation value to adjust the calculation of the Hamming distance to reflect the impact of the channel on the signal. This embodiment can select the Hamming distance or the Euclidean distance according to the specific implementation situation. The fiber channel parameter estimation value (such as the signal The Hamming distance is weighted based on channel characteristics (such as the signal-to-noise ratio and dispersion). This weighted Hamming distance reflects the impact of channel characteristics on signal transmission, making the calculation of the path metric more consistent with actual channel conditions. For example, the error probability of each bit can be weighted according to the channel noise level. For each possible path (i.e., the transition from the current state to the next state), the weighted Hamming distance between the original symbol estimate sequence and the expected symbol sequence is calculated on the state trellis. The weighted Hamming distance is used as the path metric to evaluate the quality of each path. The smaller the path metric, the more likely the path is the correct path. This embodiment searches for a surviving path by backtracing the state trellis. The specific implementation steps are as follows:

[0104] An initial path metric is set for each state in the state trellis graph. For each time step, the sequence and fiber channel parameters are estimated based on the original symbols, the path metric of each state is updated, the path with the minimum transition path metric is retained, and other paths are discarded. By backtracing the trellis graph, the path with the minimum cumulative path metric from the initial state to the final state is found, that is, the survivor path.

[0105] This embodiment extracts symbol decisions corresponding to each time step based on the survivor path. These symbol decisions are the output bits corresponding to the states passed along the survivor path from the initial state to the final state. The extracted symbol decisions are arranged in sequence to form a final symbol decision sequence. This embodiment uses fiber channel parameter estimates to optimize the decoding process, which can optimize the original symbol estimation sequence and recover a more accurate distribution communication recovery signal.

[0106] An embodiment of the present invention provides a data recovery method for a power distribution communication network based on an optical fiber network. The method obtains signal waveform data of a receiving end on an optical fiber communication link and performs feature extraction on the signal waveform data of the receiving end to obtain signal waveform distortion characteristics; minimizes inherent inter-symbol interference caused by channel characteristics as an optimization goal, performs global optimization on dispersion equalization parameters according to the signal waveform distortion characteristics, and obtains dispersion equalization pre-compensation parameters; uses the dispersion equalization pre-compensation parameters to compensate for the dispersion effect in the optical fiber communication link, and uses an adaptive filtering method based on blind equalization in the optical fiber communication link to perform blind equalization processing on the dynamic inter-symbol interference signal generated by the mismatch of the dispersion equalization pre-compensation parameters to obtain an equalized recovery signal; uses a maximum likelihood estimation algorithm to perform parameter estimation on the equalized recovery signal to obtain an original symbol estimation sequence; establishes a two-way channel transmission model based on optical fiber dispersion and multipath effects, and obtains optical fiber channel parameter estimation values ​​through learning the two-way channel transmission model; uses the optical fiber channel parameter estimation values ​​to optimize the original symbol estimation sequence to recover the power distribution communication recovery signal. Compared with the existing technology, this method addresses the problems of signal waveform distortion and inter-symbol interference caused by the superposition of the dispersion effect of the optical fiber link and the multipath effect of the distribution line. Through technologies such as the dual-channel transmission model and adaptive equalization, it effectively solves the inter-symbol interference problem caused by the mismatch between dispersion compensation and equalizer parameters, thereby greatly improving the accuracy of data recovery in the distribution communication network, reducing the power consumption and complexity of the system while ensuring the quality of data recovery, and ensuring the stable operation of the system and reliable data transmission.

[0107] It should be noted that the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of this application.

[0108] In one embodiment, Figure 2 As shown, an embodiment of the present invention provides a data recovery system for a power distribution communication network based on an optical fiber network, the system comprising:

[0109] The feature extraction module 101 is used to obtain the receiving end signal waveform data on the optical fiber communication link, and perform feature extraction on the receiving end signal waveform data to obtain signal waveform distortion features;

[0110] The compensation analysis module 102 is configured to perform global optimization of dispersion equalization parameters based on the signal waveform distortion characteristics with the optimization goal of minimizing the inherent inter-symbol interference caused by the channel characteristics to obtain dispersion equalization pre-compensation parameters;

[0111] a compensation optimization module 103 for compensating for the dispersion effect in the optical fiber communication link using the dispersion equalization pre-compensation parameters, and performing blind equalization processing on a dynamic inter-symbol interference signal generated by a mismatch in the dispersion equalization pre-compensation parameters in the optical fiber communication link using an adaptive filtering method based on blind equalization to obtain an equalized recovery signal;

[0112] The parameter estimation module 104 is configured to perform parameter estimation on the equalized restored signal using a maximum likelihood estimation algorithm to obtain an original symbol estimation sequence;

[0113] A channel analysis module 105 is configured to establish a two-way channel transmission model based on fiber dispersion and multipath effects, and obtain fiber channel parameter estimates through learning the two-way channel transmission model;

[0114] The data recovery module 106 is configured to optimize the original symbol estimation sequence using the fiber channel parameter estimation value to recover the power distribution communication recovery signal.

[0115] For the specific definition of a power distribution communication network data recovery system based on an optical fiber network, please refer to the above-mentioned definition of a power distribution communication network data recovery method based on an optical fiber network, which will not be repeated here. A person of ordinary skill in the art will appreciate that the various modules and steps described in conjunction with the embodiments disclosed in this application can be implemented in hardware, software, or a combination of both. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0116] An embodiment of the present invention provides a data recovery system for a power distribution communication network based on an optical fiber network. The system obtains receiving-end signal waveform data on an optical fiber communication link through a feature extraction module, and performs feature extraction on the receiving-end signal waveform data to obtain signal waveform distortion characteristics. A compensation analysis module takes minimizing inherent inter-symbol interference caused by channel characteristics as an optimization goal, and globally optimizes dispersion equalization parameters according to the signal waveform distortion characteristics to obtain dispersion equalization pre-compensation parameters. A compensation optimization module uses the dispersion equalization pre-compensation parameters to compensate for the dispersion effect in the optical fiber communication link, and uses an adaptive filtering method based on blind equalization in the optical fiber communication link to perform blind equalization processing on dynamic inter-symbol interference signals generated by mismatch of dispersion equalization pre-compensation parameters to obtain an equalized recovery signal. A parameter estimation module uses a maximum likelihood estimation algorithm to estimate the parameters of the equalized recovery signal to obtain an original symbol estimation sequence. A channel analysis module establishes a two-way channel transmission model based on optical fiber dispersion and multipath effects, and obtains optical fiber channel parameter estimation values ​​through learning the two-way channel transmission model. A data recovery module uses the optical fiber channel parameter estimation values ​​to optimize the original symbol estimation sequence to recover the power distribution communication recovery signal. Compared with existing technologies, this system faces the problems of signal waveform distortion and inter-symbol interference caused by the superposition of the dispersion effect of the optical fiber link and the multipath effect of the distribution line. Through technologies such as dual-path channel transmission model and adaptive equalization, it effectively solves the inter-symbol interference problem caused by the mismatch between dispersion compensation and equalizer parameters, thereby greatly improving the accuracy of data recovery in the distribution communication network, reducing the power consumption and complexity of the system while ensuring the quality of data recovery, and ensuring the stable operation of the system and reliable data transmission.

[0117] The above-described embodiments merely represent several preferred implementations of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that a person skilled in the art could make several improvements and substitutions without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be based on the scope of protection of the claims.

Claims

1. A data recovery method for a power distribution communication network based on an optical fiber network, characterized in that: The following steps are involved: Acquire receiving end signal waveform data on the optical fiber communication link, and perform feature extraction on the receiving end signal waveform data to obtain signal waveform distortion features; Taking minimizing the inherent inter-symbol interference caused by the channel characteristics as the optimization goal, the dispersion equalization parameters are globally optimized according to the signal waveform distortion characteristics to obtain the dispersion equalization pre-compensation parameters; The dispersion equalization pre-compensation parameters are used to compensate for the dispersion effect in the optical fiber communication link, and in the optical fiber communication link, a dynamic inter-symbol interference signal generated by the mismatch of the dispersion equalization pre-compensation parameters is blindly equalized using an adaptive filtering method based on blind equalization to obtain an equalized recovery signal; Performing parameter estimation on the equalized restored signal using a maximum likelihood estimation algorithm to obtain an original symbol estimation sequence; Establishing a two-way channel transmission model based on optical fiber dispersion and multipath effects, and obtaining optical fiber channel parameter estimation values ​​through learning the two-way channel transmission model; The optical fiber channel parameter estimation value is used to optimize the original symbol estimation sequence to restore the power distribution communication restoration signal.

2. A method for recovering data of a power distribution communication network based on an optical fiber network as claimed in claim 1, characterized in that: The step of taking minimization of the inherent inter-symbol interference caused by the channel characteristics as the optimization goal and performing global optimization on the dispersion equalization parameters according to the signal waveform distortion characteristics to obtain the dispersion equalization pre-compensation parameters comprises: The support vector machine algorithm is used to learn the signal waveform distortion characteristics and predict the fiber dispersion coefficient and power line multipath delay spread data of the current communication link; According to the fiber dispersion coefficient and power line multipath delay spread data, the current communication link distortion degree is quantified; When the distortion degree of the current communication link exceeds the preset distortion threshold, the dispersion equalization parameters are globally optimized using a genetic algorithm with the goal of minimizing the inherent inter-symbol interference caused by channel characteristics to obtain the dispersion equalization pre-compensation parameters.

3. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 2, characterized in that: The step of using a genetic algorithm to globally optimize the dispersion equalization parameters comprises: A set of random dispersion equalization parameter combinations are generated as an initial population; the dispersion equalization parameters include dispersion compensation depth and equalizer tap coefficients; The optimization goal is to minimize the inherent inter-symbol interference caused by channel characteristics, and the dispersion equalization parameters are used as the search space of the genetic algorithm; In each iteration, the cost function value corresponding to each dispersion equalization parameter combination is calculated to obtain the fitness; According to the fitness, crossover and mutation operations are performed, and multiple rounds of iterative updates are performed until the preset convergence conditions are met, and the optimal dispersion equalization parameter combination is obtained, and the optimal dispersion equalization parameter combination is used as the dispersion equalization pre-compensation parameter; the dispersion equalization pre-compensation parameter includes a dispersion compensation depth value and an equalizer tap coefficient value.

4. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 1, characterized in that: The step of using the adaptive filtering method based on blind equalization to perform blind equalization processing on the dynamic inter-symbol interference signal generated by the mismatch of dispersion equalization pre-compensation parameters to obtain the equalized recovery signal includes: Receiving a dynamic inter-symbol interference signal generated by a dispersion equalization pre-compensation parameter mismatch, and calculating a second-order moment and a fourth-order moment of the dynamic inter-symbol interference signal to obtain a statistical feature of the interference signal; The autocorrelation function of the interference signal is calculated according to its statistical characteristics, and the signal delay characteristics are estimated according to the peak position of the autocorrelation function; The tap coefficients required by the blind equalizer are set according to the signal delay characteristics, and the tap coefficients are iteratively updated using a blind equalization algorithm; In each iterative update, the minimum distance from the signal point on the constellation diagram of the current blind equalizer output signal to the boundary of the decision domain is calculated to obtain the distortion; If the fluctuation of the distortion degree for multiple consecutive periods is less than the preset fluctuation threshold and the distortion degree is lower than the preset first distortion threshold, it is determined that the blind equalizer has reached a stable state, the tap coefficient of the current blind equalizer is locked, and the current equalization state is maintained; After the tap coefficients of the current blind equalizer converge and are locked, the dynamic inter-symbol interference signal is filtered using the current blind equalizer according to the locked tap coefficients, and an equalization recovery signal is output.

5. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 4, characterized in that: If the distortion exceeds a preset second distortion threshold or the constellation diagram deflection angle exceeds a preset deflection angle threshold, the locking state of the current blind equalizer tap coefficient is released, and the tap coefficient is iteratively updated again.

6. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 4, characterized in that: The update step size of the current blind equalizer tap coefficients is adjusted according to the distortion.

7. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 1, characterized in that: The step of obtaining the estimated value of the optical fiber channel parameter by learning the two-way channel transmission model comprises: According to the two-way channel transmission model, a two-way channel transmission process is simulated by using a step-by-step Fourier method to obtain channel state data in the optical fiber communication link; Extracting optical fiber channel characteristic parameters according to the channel status data in the optical fiber communication link; the optical fiber channel characteristic parameters include optical fiber dispersion coefficient, group delay, multipath amplitude and multipath delay; A recurrent neural network is used to perform nonlinear mapping on the characteristic parameters of the optical fiber channel to obtain an estimated value of the optical fiber channel parameter.

8. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 1, characterized in that: The step of optimizing the original symbol estimation sequence by using the fiber channel parameter estimation value to recover the power distribution communication recovery signal comprises: Introducing the fiber channel parameter estimation value, using the Viterbi decoding algorithm to decode the original symbol estimation sequence to obtain a symbol decision sequence; Encoding the symbol decision sequence to obtain a coded data sequence, and segmenting the coded data sequence to obtain error correction data blocks; Obtaining the average burst error length and frequency of the channel according to the estimated values ​​of the optical fiber channel parameters, and determining the interleaving depth according to the average burst error length and frequency of the channel; constructing an interleaving matrix according to the interleaving depth, and rearranging the error correction data block using the interleaving matrix to form an interleaved data block; The interleaved data block is decoded using a decoding algorithm based on an error correction code to recover a power distribution communication recovery signal.

9. A method for recovering data in a power distribution communication network based on an optical fiber network as claimed in claim 8, characterized in that: The step of introducing the fiber channel parameter estimation value and decoding the original symbol estimation sequence using the Viterbi decoding algorithm to obtain the symbol decision sequence comprises: Constructing a state trellis diagram according to the constraint length and coding structure of the convolutional code encoder at the transmitting end; each state in the state trellis diagram represents a coding state of the convolutional code encoder; In the decoding process, according to the original symbol estimation sequence, the Hamming distance is weighted by the fiber channel parameter estimation value, and the path metric of each state in the state trellis diagram transferring to the next state is calculated by using the weighted Hamming distance; By backtracking the state lattice graph, a path with a minimum cumulative path metric from an initial state to a final state is found to obtain a surviving path; The symbol decisions corresponding to each coding state on the surviving path are arranged in sequence to form a symbol decision sequence.

10. A data recovery system for power distribution communication network based on optical fiber network, characterized in that: The system comprises: A feature extraction module is used to obtain the receiving end signal waveform data on the optical fiber communication link, and perform feature extraction on the receiving end signal waveform data to obtain signal waveform distortion features; A compensation analysis module, configured to globally optimize the dispersion equalization parameters according to the signal waveform distortion characteristics with minimizing the inherent inter-symbol interference caused by the channel characteristics as the optimization goal, and obtain the dispersion equalization pre-compensation parameters; A compensation optimization module, used to compensate for the dispersion effect in the optical fiber communication link by using the dispersion equalization pre-compensation parameters, and in the optical fiber communication link, perform blind equalization processing on the dynamic inter-symbol interference signal generated by the mismatch of the dispersion equalization pre-compensation parameters by using an adaptive filtering method based on blind equalization to obtain an equalized recovery signal; A parameter estimation module, used to perform parameter estimation on the equalized recovery signal using a maximum likelihood estimation algorithm to obtain an original symbol estimation sequence; A channel analysis module is used to establish a two-way channel transmission model based on optical fiber dispersion and multipath effects, and obtain optical fiber channel parameter estimation values ​​through learning the two-way channel transmission model; The data recovery module is used to optimize the original symbol estimation sequence by using the fiber channel parameter estimation value to recover the power distribution communication recovery signal.

Images