A single fiber bidirectional wavelength division transmission method for a passive optical network

Abstract

The application relates to the technical field of optical fiber communication, and particularly discloses a single-fiber bidirectional wavelength division transmission method for a passive optical network, which comprises the following steps: after protocol shunting and service level baseline parameter extraction are performed on access side multi-protocol data streams, wavelength adaptive pre-distortion conversion is performed on each channel based on a compensation vector returned by a receiving end, a pre-compensation optical signal and an out-of-band pilot signal of a non-service wavelength band are coupled and superimposed into a composite optical flow carrying telemetry, after transmission through a double-redundancy optical fiber network, a single-wave channel state feature is extracted in a non-invasive manner by intercepting and sampling the optical flow at the receiving end, finally, a service signal is output through wave demodulation, a compensation vector is calculated based on an optical fiber state matrix, and the compensation vector is returned to the sending end through a reverse optical path closed loop. The scheme sinks the monitoring granularity from the total power level to the single wavelength level with the aid of out-of-band telemetry, and continuously iterates the closed loop pre-compensation adaptive repair of the channel degradation at the receiving and sending ends, effectively breaks through the residual error bottleneck, and guarantees the reliable transmission of power grid protection services.

Description

Technical Field

[0001] This application relates to the field of optical fiber communication technology, and more specifically, to a single-fiber bidirectional wavelength division transmission method for passive optical networks. Background Technology

[0002] With the continuous advancement of power grid digitalization and intelligentization, the types of services carried by power communication networks are becoming increasingly diversified. Critical data streams such as tripping commands, differential protection sampling messages, and substation video monitoring place extremely stringent requirements on transmission bandwidth and reliability. However, the resources of the early-built power communication fiber optic networks are nearing saturation, and the contradiction between the explosive growth of business volume and the physical capacity of fiber optics is becoming increasingly acute. Relying on laying new fiber optic cables or leasing optical cables from operators will face practical difficulties such as long approval cycles, high construction difficulty, and a significant increase in operating costs.

[0003] Existing single-fiber bidirectional wavelength division multiplexing (WDM) transmission systems typically employ optical line protection mechanisms to ensure link reliability. The core idea is to monitor the overall optical power of the composite optical signal transmitted on the backbone fiber, triggering a primary / backup link switchover upon detecting a power drop. However, this macroscopic total power monitoring granularity suffers from a significant scale mismatch with the up to sixteen or more independent wavelength channels in dense wavelength division multiplexing (DWDM) systems. When a specific wavelength carrying critical protection commands experiences severe power fading due to laser aging, local four-wave mixing, or dispersion accumulation, the remaining wavelengths remain at normal levels. This results in the total optical power change being far below the switching threshold, making it impossible for the system to detect the minute degradation of that single channel, leading to silent packet loss for critical services. Furthermore, even if a channel-by-channel pre-compensation mechanism is introduced at the transmitting end to correct the aforementioned single-channel degradation, the existing scheme still treats each wavelength channel as an independent physical entity for isolated compensation, completely ignoring the coupling interference caused by Kerr nonlinearity in the co-propagation of multiple high-power optical signals in the same fiber core, such as inter-channel cross-phase modulation and four-wave mixing. This makes it impossible for the compensation vector to converge to the global optimum in the closed-loop iteration, and the residual bit error rate of the system is constrained by an uneliminable nonlinear floor, which seriously threatens the reliable transmission of power grid protection commands on long-distance optical fiber links.

[0004] Therefore, an optimized single-fiber bidirectional wavelength division multiplexing (WDM) transmission scheme for passive optical networks is desired. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks, comprising: Step 1: Perform protocol splitting and parsing and service level baseline parameter extraction on the multi-protocol data stream collected from the access side, which includes tripping command stream, high-frequency sampling message and video frame sequence, to obtain the service load matrix and driving parameter set; Step 2: Based on the driving parameter set and the compensation vector returned by the closed loop at the receiver, perform wavelength adaptive pre-distortion transformation on the service payload matrix to obtain an independent set of pre-compensated optical signals for each channel. Step 3: Perform multi-wavelength multiplexing on the pre-compensated optical signal set in the pure optical physical layer, and couple and superimpose the multiplexing result with the out-of-band pilot signal generated independently in the non-service band to generate a telemetry composite optical flow for single-fiber bidirectional transmission. Step 4: Perform all-optical link dual-redundancy transmission on the telemetry composite optical flow, and at the receiving end, use a bypass splitter to non-intrusively intercept the sampled optical flow for single-channel state feature extraction to obtain the arrival state composite optical flow and fiber state matrix. Step 5: Demodulate the arrival state composite optical flow to output the service signal, and at the same time perform single-wavelength deviation calculation on the fiber state matrix to generate a compensation vector, and then transmit the compensation vector back to Step 2 via the reverse optical path closed loop.

[0006] Compared with existing technologies, this application proposes a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks. It performs protocol splitting and parsing, and extracts service level baseline parameters from a multi-protocol data stream containing tripping command streams, high-frequency sampling messages, and video frame sequences on the access side. Based on the obtained driving parameter set and the compensation vector returned from the receiver's closed loop, it performs wavelength adaptive pre-distortion conversion on the service payload matrix. At the pure optical physical layer, it performs multi-wavelength multiplexing and couples it with out-of-band pilot signals from non-service bands to generate a telemetry-carrying composite optical stream. After parallel transmission via a dual-redundant all-optical link, the receiver non-intrusively intercepts the sampled optical stream to extract single-channel state features. While demodulating the arrival state composite optical stream to output the service signal, it simultaneously performs single-wavelength deviation calculation on the fiber state matrix to generate a compensation vector, which is then returned to the transmitter via a reverse optical path in a closed loop. This constructs an end-to-end adaptive transmission system from transmitter pre-distortion to receiver state perception and then to reverse closed-loop correction. It achieves precise repair of single-wavelength-level degradation without interrupting services, ensuring highly reliable transmission of critical power grid protection services. Attached Figure Description

[0007] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0008] Figure 1 This is a flowchart of a single-fiber bidirectional wavelength division transmission method for a passive optical network according to an embodiment of this application; Figure 2This is a schematic diagram of data flow in a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for a passive optical network according to an embodiment of this application. Figure 3 This is a flowchart illustrating a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. The method involves performing wavelength adaptive pre-distortion transformation on the service payload matrix to obtain an independent set of pre-compensated optical signals for each channel, based on a set of driving parameters and a compensation vector returned from the closed loop at the receiving end. Figure 4 This is a flowchart illustrating the coupled sensing joint pre-distortion waveform reconstruction of a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. Figure 5 This document describes a flowchart of a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. The method involves performing all-optical link dual-redundancy transmission of the telemetry composite optical flow and, at the receiving end, using a bypass splitter to non-intrusively intercept the sampled optical flow for single-channel state feature extraction, in order to obtain the arrival state composite optical flow and the fiber state matrix. Figure 6 This is a flowchart illustrating a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. The method involves performing wavelength division demodulation on the arriving state composite optical flow to output a service signal, simultaneously performing single-wavelength deviation calculation on the fiber state matrix to generate a compensation vector, and then transmitting the compensation vector back to step two via a closed-loop reverse optical path. Detailed Implementation

[0009] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0010] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0011] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0012] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0013] Currently, single-fiber bidirectional wavelength division multiplexing (WDM) transmission systems generally rely on macroscopic monitoring of the overall optical power of composite optical signals and primary / backup link switching mechanisms to ensure link reliability. However, dense WDM systems often carry more than ten independent wavelength channels in parallel. When a specific wavelength carrying critical protection commands suffers severe degradation due to device aging or the accumulation of nonlinear effects, the total power change may be far below the switching threshold. The system cannot detect the slight fading of this single channel, leading to silent packet loss of critical services. Meanwhile, even with the introduction of channel-by-channel pre-compensation, existing solutions still treat each wavelength channel as an independent physical entity for isolated correction, failing to consider inter-channel coupling interference caused by Kerr nonlinearity when multiple optical signals are co-transmitted on the same fiber core. This results in the compensation iteration failing to converge to the global optimum, and the system's residual bit error rate being constrained by the nonlinear floor. Therefore, this application proposes a single-fiber bidirectional WDM transmission method for passive optical networks. This method first performs protocol splitting and parsing and service level baseline parameter extraction on the access side for multi-protocol data streams containing tripping commands, high-frequency sampling messages, and video frame sequences to obtain the service payload matrix and driving parameter set. Then, based on the driving parameter set and the compensation vector transmitted back through the closed loop at the receiver, wavelength adaptive pre-distortion transformation is performed on the service payload matrix to generate a pre-compensated optical signal set. Next, multi-wavelength multiplexing is performed at the pure optical physical layer, and the multiplexing result is coupled and superimposed with out-of-band pilot signals independently generated in non-service bands to form a telemetry composite optical flow. This composite optical flow is then transmitted through a dual-redundant fiber optic network. After all-optical transmission, the receiver uses a bypass splitter to non-intrusively intercept the sampled optical flow. With the help of pilot baseline anchors, it completes the physical attenuation and phase shift jitter quantization of a single channel to form an optical fiber state matrix. Finally, while demodulating the arrival state composite optical flow to output the service signal, it performs single-wavelength deviation calculation based on the optical fiber state matrix to generate a compensation vector, which is then transmitted back to the transmitter via the reverse optical path for use in the next cycle of pre-distortion. This allows the system to continuously and adaptively repair single-channel degradation without interrupting the service, effectively breaking through the residual error bottleneck and ensuring the reliable transmission of critical power grid protection services.

[0014] Figure 1 This is a flowchart of a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for a passive optical network according to an embodiment of this application. Figure 2 This is a schematic diagram of data flow in a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for a passive optical network according to an embodiment of this application. Figure 1 and Figure 2 As shown, a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application includes: S1, performing protocol splitting and parsing and service level baseline parameter extraction on a multi-protocol data stream collected from the access side, including a tripping command stream, high-frequency sampling messages, and video frame sequences, to obtain a service payload matrix and a driving parameter set; S2, performing wavelength adaptive pre-distortion transformation on the service payload matrix based on the driving parameter set and the compensation vector returned from the closed loop at the receiving end to obtain an independent pre-compensated optical signal set for each channel; S3, performing multi-wavelength multiplexing processing on the pre-compensated optical signal set at a pure optical physical layer. The combined optical stream is coupled and superimposed with the out-of-band pilot signal generated independently in the non-service band to generate a telemetry-carrying composite optical stream for bidirectional transmission on a single fiber; S4, the telemetry-carrying composite optical stream is transmitted with dual redundancy in an all-optical link, and at the receiving end, a bypass splitter is used to non-intrusively intercept the sampled optical stream for single-channel state feature extraction to obtain the arrival state composite optical stream and the fiber state matrix; S5, the arrival state composite optical stream is demodulated to output the service signal, and at the same time, a single-wavelength deviation calculation is performed on the fiber state matrix to generate a compensation vector, and the compensation vector is returned to step two via a closed loop through the reverse optical path.

[0015] Specifically, in step S1, the multi-protocol data stream collected from the access side, including tripping command streams, high-frequency sampling messages, and video frame sequences, undergoes protocol splitting and parsing, and service level baseline parameter extraction to obtain the service payload matrix and driving parameter set. It should be noted that because the data stream collected from the access side contains tripping command streams, high-frequency sampling messages, and video frame sequences following different protocol specifications, the requirements for bit error rate tolerance and transmission delay vary by magnitude among different services. If it directly enters the wavelength conversion stage without classification and parsing, it will be impossible to establish an optical layer transmission reference matching the service level for each channel. Based on this, the technical solution of this application first performs protocol splitting and parsing, and service level baseline parameter extraction on the multi-protocol data stream collected from the access side to obtain the service payload matrix and driving parameter set. Through the above processing, the mixed multi-protocol data stream can be transformed into structured payloads and accurate optical layer driving parameters corresponding one-to-one with the wavelength division multiplexing channels, providing a data foundation and physical reference for subsequent channel-by-channel pre-distortion conversion.

[0016] More specifically, in a concrete example of this application, the multi-protocol data streams aggregated on the access side are first segmented and priority fields are parsed. Then, multi-layer protocol header stripping and two-dimensional payload reassembly are performed on the segmented packets. Finally, the baseline transmit power calculation and center wavelength standard value binding for each channel are completed based on the service level identifier. During the protocol segmentation and priority field parsing stage, the data link layer packet header of the multi-protocol data streams aggregated on the access side is read bit by bit. Based on the Ethernet type field, the GOOSE tripping command stream, SV high-frequency sampling packets, and H.265 video frame sequences are segmented and separated into independent classified packet sequences. Simultaneously, the IEEE 802.1p priority bits are extracted from the virtual LAN label and quantized and mapped to the corresponding service level identifier. The tripping command stream is mapped to the highest priority, the SV high-frequency sampling packets to the medium priority, and the video frame sequences to the lower priority. The classified packet sequences are then passed to the subsequent payload stripping stage, and the service level identifier is bypassed and output to the baseline parameter calculation stage.

[0017] After the protocol-level traffic splitting is completed, the next stage is multi-layer protocol header stripping and two-dimensional payload reassembly. The classified message sequence output from the previous stage is de-encapsulated frame by frame, sequentially stripping the protocol header fields of the MAC layer, IP layer, and TCP / UDP transport layer, retaining only the pure data segments carrying the actual service content. Based on this, according to the wavelength division multiplexing channel allocation rules, the extracted multiple pure data segments are reassembled and mapped in two dimensions according to the time slice dimension and the spatial channel dimension, generating a service payload matrix where row indices correspond to time slices and column indices correspond to spatial channels.

[0018] After the service payload matrix is ​​constructed, the stage of calculating the reference transmit power and binding it to the center wavelength standard value begins. The priority weights of each channel in the aforementioned service level identifiers are read, and combined with the reference minimum optical power of the receiver's photodetector and the power increment step size, the reference transmit power of the k-th channel is calculated. The reference transmit power of the k-th channel... Equal to the system reference minimum optical power Add the channel priority weight With power increment step size The product of these factors, minus the bit error rate threshold tolerable by the channel. The calculated redundancy loss compensation term is: in, The minimum reference optical power that the photodetector at the receiving end can stably demodulate, expressed in dBm. Let be the priority weight of the k-th channel, and be a dimensionless constant. This represents the power increment step size, in dB. This represents the maximum bit error rate threshold that the k-th channel can tolerate in the service level agreement. The reference transmit optical power calculated for each channel is bound to the pre-allocated center wavelength standard value in the ITU-T standard wavelength grid, and encapsulated to generate a set of driving parameters.

[0019] Specifically, in step S2, based on the driving parameter set and the compensation vector returned from the closed loop at the receiving end, wavelength adaptive pre-distortion conversion is performed on the service payload matrix to obtain an independent pre-compensated optical signal set for each channel. It should be noted that, given that the driving parameter set generated in step one only reflects the reference transmission state of each channel under factory calibration or initial configuration conditions, while actual fiber optic links continuously deteriorate due to factors such as device aging, environmental temperature drift, and the accumulation of nonlinear effects, if the optical wavelength conversion board is directly driven to emit light using fixed reference parameters, the actual arriving power and phase of each channel will deviate from the expected values, failing to meet the critical service requirements of low bit error rate (BER) for tripping command streams and other critical services. The application faces stringent requirements in terms of scale. Therefore, the technical solution of this application further performs wavelength adaptive pre-distortion transformation on the service payload matrix based on the driving parameter set and the compensation vector returned from the closed loop at the receiver to obtain an independent set of pre-compensated optical signals for each channel. Through the above processing, power and phase pre-compensation can be performed for the real-time degradation state of each wavelength channel before the optical signal is injected into the optical fiber, ensuring that the optical signal of each channel can still maintain effective demodulation margin at the receiver after long-distance transmission fading.

[0020] Figure 3 This is a flowchart illustrating a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. The method involves performing wavelength adaptive pre-distortion transformation on the service payload matrix based on a set of driving parameters and a compensation vector returned via closed-loop transmission from the receiver to obtain an independent set of pre-compensated optical signals for each channel. Figure 3 As shown, step S2 includes: S21, using the driving parameter set as a reference tensor, and performing channel-by-channel tensor superposition with the power compensation components of each channel in the compensation vector to obtain the corrected driving parameters; S22, based on the corrected driving parameters, performing electrical domain pre-distortion waveform reconstruction and digital-to-analog conversion processing on the service payload matrix to generate a pre-distorted electrical signal; S23, based on the pre-distorted electrical signal, driving the tunable laser to perform laser external modulation and photonic domain mapping to generate a pre-compensated optical signal set.

[0021] In step S21, the driving parameter set is used as a reference tensor, and channel-by-channel tensor superposition is performed with the power compensation components of each channel in the compensation vector to obtain the corrected driving parameters. It should be noted that since the reference transmit power in the driving parameter set only reflects the theoretical setting value of each channel under initial calibration conditions, while the compensation vector carries channel-by-channel power compensation components and phase correction values ​​calculated by the receiver based on the actual physical state of the optical fiber, the two must be fused before emission to form an effective driving command reflecting the real-time degradation state of the link. Based on this, the technical solution of this application further uses the driving parameter set as a reference tensor, and channel-by-channel tensor superposition is performed with the power compensation components of each channel in the compensation vector to obtain the corrected driving parameters. Through the above processing, the initial baseline setting and the real-time correction values ​​of the closed-loop feedback can be unified and integrated into a set of correction parameters that can directly drive subsequent electrical domain pre-distortion and laser modulation.

[0022] More specifically, in a concrete example of this application, the reference transmitted optical power of the k-th channel in the driving parameter set and the power compensation component of the corresponding channel in the compensation vector are read, and a channel-by-channel tensor superposition operation is performed on the two. The corrected driving power of the k-th channel... Equal to the reference transmit optical power of this channel Add laser aging penalty factor With the channel power compensation component The product of, i.e.: in, The k-th channel reference transmit optical power calculated in step one, in dBm. The laser aging penalty factor, used to offset the decline in luminous efficiency caused by long-term operation, is a dimensionless constant not less than 1. This is the power compensation component of the k-th channel in the compensation vector, in dB, generated by the receiver based on the fiber state matrix. Taking a high-priority channel carrying a tripping command stream as an example, if the reference transmit optical power of this channel is 3 dBm, the power compensation component fed back by the compensation vector is 1.5 dB, and the aging penalty factor is 1.2, then the corrected drive power is calculated to be 4.8 dBm, which is 1.8 dB higher than the reference value to offset the actual fading of the current link. After completing the power correction calculation for all effective channels, the phase correction amount and center wavelength fine-tuning offset of each channel in the compensation vector are extracted simultaneously. The corrected power, phase, and wavelength parameters are uniformly encapsulated according to the channel index to generate the corrected drive parameters, which are then transferred to the subsequent electrical domain pre-distortion waveform reconstruction stage.

[0023] In step S22, based on the corrected driving parameters, the service payload matrix undergoes electrical domain pre-distortion waveform reconstruction and digital-to-analog conversion to generate a pre-distortion electrical signal. It should be noted that since the corrected driving parameters already incorporate the real-time correction values ​​from the reference setting and closed-loop feedback, they need to be further converted into a voltage waveform capable of driving the laser modulator to pre-counteract fiber optic link fading in the electrical domain. Therefore, the technical solution of this application further performs electrical domain pre-distortion waveform reconstruction and digital-to-analog conversion on the service payload matrix based on the corrected driving parameters to generate a pre-distortion electrical signal. Through the above processing, reverse pre-compensation for the fiber optic fading characteristics of each channel can be completed in the electrical domain stage, enabling the subsequent laser modulation output optical signal to carry power and phase characteristics that counteract link damage.

[0024] More specifically, in a particular example of this application, the voltage gain coefficient of the k-th channel is resolved from the modified driving parameters. With phase compensation bias Pre-distortion waveform reconstruction is performed on the instantaneous bit values ​​of this channel in the service payload matrix. The instantaneous voltage of the pre-distorted electrical signal of the k-th channel at time t is... Equal to voltage gain coefficient The corresponding bit values ​​in the service payload matrix With complex exponential phase compensation factor The product of the products, i.e.: in, The voltage gain coefficient is obtained from the modified drive power, and it is a dimensionless proportionality coefficient. Let be the instantaneous bit value of the k-th channel in the traffic payload matrix at time t. The self-channel phase compensation bias, measured in radians, is used to compensate for fiber dispersion and nonlinear phase shift. The above calculations are performed independently for each channel, and the gain coefficient and phase bias used for the pre-distortion calculation of each channel are strictly bound to the physical state of that channel itself. After digital-to-analog conversion, the pre-distorted voltage waveforms of each channel are unified and integrated into a pre-distorted electrical signal.

[0025] In another specific example of this application, considering that dense wavelength division multiplexing transmits up to 16 different wavelength optical signals in parallel within the same fiber core, with each signal sharing the same fiber medium, and the silica material of the fiber core having an inherent third-order nonlinear polarizability, when multiple optical signals propagate simultaneously in the fiber core, the Kerr nonlinear effect will establish physical coupling channels between the channels, manifesting as two mechanisms: cross-phase modulation and four-wave mixing. The aforementioned channel-by-channel independent pre-distortion scheme does not pre-cancel this inter-channel coupling relationship, and the residual phase error accumulated after long-distance transmission will prevent the bit error rate from being further reduced. Therefore, an inter-channel nonlinear coupling sensing mechanism is introduced in the electrical domain pre-distortion stage, sequentially performing coupling weight matrix construction, cross-channel interference phase prediction, and coupled sensing joint waveform reconstruction.

[0026] Figure 4 This is a flowchart illustrating the coupled sensing joint pre-distortion waveform reconstruction of a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. Figure 4 As shown, step S22 includes: S221, determining the channel coupling weight matrix based on the modified driving parameters; S222, performing cross-channel nonlinear interference phase prediction on the service payload matrix based on the channel coupling weight matrix to obtain the cross-channel interference phase vector; S223, performing coupled sensing joint pre-distortion waveform reconstruction on the service payload matrix based on the cross-channel interference phase vector and the modified driving parameters to obtain the pre-distortion electrical signal.

[0027] In step S221, the channel coupling weight matrix is ​​determined based on the modified driving parameters. It should be noted that, given that when multiple optical signals propagate together within the same fiber core, the nonlinear coupling strength between any two channels of different wavelengths is not uniform and constant, but depends simultaneously on their respective transmit power levels and the frequency spacing between them. Higher power and smaller spacing result in more intense coupling. If this differentiated coupling relationship is not quantitatively modeled before waveform reconstruction, subsequent interference prediction will lack a physical basis. Therefore, the technical solution of this application further determines the channel coupling weight matrix based on the modified driving parameters. Through the above processing, the three independent physical dimensions of the fiber physical layer—the nonlinear refractive index parameter, the channel transmit power, and the frequency spacing—can be integrated into a unified, calculable weight matrix, providing a numerical basis consistent with the actual physical behavior of the optical fiber for subsequent channel-by-channel interference prediction.

[0028] More specifically, in a concrete example of this application, the modified drive power and center frequency of each channel are analytically derived from the modified drive parameters. Using fiber Kerr nonlinearity theory, the nonlinear coupling weight between any two channels k and m is analytically calculated. The coupling weight of channel m to channel k is equal to the product of the fiber nonlinear refractive index coefficient and the effective interaction length, divided by the effective mode field area, multiplied by the product of the linear values ​​of the modified drive powers of the two channels, and divided by the sum of the square of the center frequency spacing between the two channels and the regularization constant, i.e.: in The nonlinear coupling weight applied by channel m to channel k, in rad / mW. The nonlinear refractive index coefficient of the optical fiber medium, in units of m. 2 / W, To comprehensively consider the effective nonlinear interaction length after fiber attenuation, the unit is km. The effective mode area of ​​the optical fiber, in μm. 2 , and These are the linear domain optical power values ​​of the channel-corrected driving power for the k-th and m-th channels, respectively, after being restored from the logarithmic domain, in mW. The center frequency distance between the two channels, in Hz. A regularization smoothing constant, in Hz, is used to prevent the denominator from becoming zero during frequency degradation. 2 After traversing all pairwise combinations of the N effective channels to complete the above calculations, all weight values ​​are assembled into an N×N symmetric matrix with its diagonals set to zero, i.e., the channel coupling weight matrix. In this formula structure, the dual-channel power product term in the numerator naturally results in a larger coupling weight between high-power channels, while the frequency spacing square term in the denominator makes the coupling strength between adjacent channels much higher than that between distant channels. This characteristic is consistent with the physical law of cross-phase modulation in Kerr nonlinearity. Taking a 16-channel dense wavelength division multiplexing channel in a power communication scenario as an example, assuming that the corrected drive power of the first channel carrying the tripping command stream is 4.8 dBm and the center frequency is 193.1 THz, the corrected drive power of the second channel carrying video surveillance is 1.5 dBm and the center frequency is 193.2 THz, and the corrected drive power of the eighth channel carrying administrative telephone is 2.0 dBm and the center frequency is 193.8 THz, and taking the fiber nonlinear refractive index coefficient as 2.6×10 -20 m 2 / W, effective interaction length of 20 km, effective model field area of ​​80 μm 2The coupling weight of the second channel to the first channel is higher due to the frequency spacing of only 100 GHz, while the coupling weight of the eighth channel to the first channel is lower due to the frequency spacing of 700 GHz. This allows for precise quantification of the different coupling strengths between the high-priority tripping command channel and the low-priority video surveillance channel, as well as between the adjacent channel and the remote channel, avoiding coarse handling of nonlinear effects.

[0029] In step S222, based on the channel coupling weight matrix, cross-channel nonlinear interference phase prediction is performed on the service payload matrix to obtain the cross-channel interference phase vector. It should be noted that since the channel coupling weight matrix characterizes the static coupling strength relationship between channels, while the actual nonlinear interference depends on the instantaneous power level of each interference source channel at each moment, the phase interference exerted on neighboring channels by the same channel when transmitting high-level bit 1 and low-level bit 0 are drastically different. If only static weights are relied upon without combining real-time envelope dynamics of the service data for time-by-time prediction, the obtained compensation amount will be a constant bias independent of the data content, unable to track rapidly changing nonlinear phase jitter. Therefore, the technical solution of this application further performs cross-channel nonlinear interference phase prediction on the service payload matrix based on the channel coupling weight matrix to obtain the cross-channel interference phase vector. Through the above processing, the inter-channel nonlinear phase interference that the optical signal will suffer during transmission can be quantized and calculated time-by-time and channel-by-channel in the digital electrical domain before the optical signal is injected into the optical fiber, providing a time-varying and accurate phase compensation target for the subsequent pre-cancellation injection.

[0030] More specifically, in a concrete example of this application, the instantaneous bit values ​​of each channel in the traffic payload matrix are used in conjunction with the weights in the channel coupling weight matrix to perform aggregated prediction of the cross-channel cumulative parasitic phase offset experienced by the k-th target channel at time t. This offset is equal to the weighted sum of the coupling weights of all other co-transmission channels and the corresponding instantaneous normalized power values ​​of the channels, i.e.: in, This represents the cumulative parasitic phase offset across channels at time t due to cross-phase modulation effects, expressed in rad. This represents the total number of valid service channels that are active in real time during wavelength division multiplexing. Let be the nonlinear coupling weights of channel m to channel k in the channel coupling weight matrix. Let t be the instantaneous normalized power value of the m-th channel in the service load matrix at time t. After traversing all N channels and completing the above aggregation calculation, the time-varying parasitic phase offset of each channel is encapsulated into a cross-channel interference phase vector. This step reverses the nonlinear phase crosstalk that originally occurred in the physical transmission process of optical fiber and advances it to the digital electrical domain of the transmitting end for accurate prediction, realizing the prediction of nonlinear optical fiber damage before transmission. Taking the aforementioned 16-channel dense wavelength division multiplexing channel as an example, at a certain moment, the first channel carrying the trip command stream is in a high-level bit state, and its instantaneous normalized power value is 1, while the second channel carrying video surveillance is simultaneously in a low-level bit state, and its instantaneous normalized power value is close to 0. At this time, the instantaneous phase interference exerted by the second channel on the first channel is close to zero. However, if the second channel switches to high-level bit transmission at the next moment, its instantaneous phase interference on the first channel will increase with the coupling weight. The numerical level rises rapidly, and this disturbance dynamic that changes rapidly with bit flips is precisely what is captured by time-by-time aggregation prediction, thus avoiding the compensation inaccuracies caused by relying solely on static constant bias.

[0031] In step S223, based on the cross-channel interference phase vector and the corrected driving parameters, the service payload matrix is ​​reconstructed using coupled sensing joint pre-distortion waveform to obtain the pre-distorted electrical signal. It should be noted that, given that the channel coupling weight matrix characterizing the static coupling strength between channels and the cross-channel interference phase vector reflecting the instantaneous envelope dynamics of the service data have been obtained through the aforementioned two sub-steps, this information needs to be finally fused with the original self-channel linear compensation parameters. Otherwise, each channel will remain in an isolated compensation state, failing to eliminate nonlinear crosstalk between channels at its source. Therefore, the technical solution of this application further reconstructs the service payload matrix using coupled sensing joint pre-distortion waveform based on the cross-channel interference phase vector and the corrected driving parameters to obtain the pre-distorted electrical signal. Through the above processing, the original paradigm of isolated compensation for each channel can be elevated to a collaborative pre-distortion paradigm of full-channel joint sensing, ensuring that the transmitted waveform of each channel contains the perception and pre-response to the real-time power state of all neighboring channels.

[0032] More specifically, in a particular example of this application, three input data streams are combined, namely the voltage gain coefficients resolved from the modified drive parameters. Phase compensation bias with self-confidence The cumulative parasitic phase offset of the k-th channel read from the inter-channel interference phase vector at time t. and instantaneous bit values ​​extracted from the service payload matrix. Perform coupled sensing-based joint pre-distortion waveform reconstruction. The instantaneous voltage of the k-th channel pre-distortion electrical signal is equal to the product of the voltage gain coefficient, the traffic load bit value, and the joint complex exponential phase factor composed of the self-channel linear phase compensation and cross-channel nonlinear interference pre-cancellation, i.e.: in, The instantaneous voltage of the pre-distorted electrical signal of the k-th channel at time t after coupled sensing and joint pre-distortion is expressed in V. Here, represents the voltage gain coefficient of the channel, and represents a dimensionless proportionality coefficient. These are the corresponding bit values ​​in the service payload matrix. This is the linear phase compensation bias of the self-channel, expressed in rad. This represents the cumulative parasitic phase offset of the k-th channel in the inter-channel interference phase vector at time t, in rad. Compare this to the formula in Example 1 above. The improved scheme adds a new term to the exponential term of the complex exponential phase factor. This crucial time-varying interchannel pre-cancellation term physically means that an anti-phase component with the opposite polarity to the predicted interchannel parasitic phase is actively injected at the transmitter. This ensures that the positive parasitic phase shift that gradually accumulates in the optical fiber due to Kerr nonlinearity is precisely canceled by the anti-phase component pre-injected at the transmitter. The two components are superimposed at the receiver and approach zero, thus eliminating inter-channel nonlinear crosstalk at its source. After traversing all N channels and completing joint reconstruction, the pre-distorted voltage waveforms of each channel are unified and integrated into a pre-distorted electrical signal, which is then output to the subsequent laser electro-optic modulation unit. Taking the aforementioned 16-channel dense wavelength division multiplexing (DWDM) channel as an example, assuming that at a certain moment, the self-channel phase compensation offset of the first channel carrying the trip command stream is 0.35 rad, and the cumulative parasitic phase offset calculated by cross-channel interference phase prediction is 0.12 rad, then the exponent term of the joint complex exponential phase factor is calculated to be 0.23 rad. That is, a -0.12 rad reverse phase component is pre-injected at the transmitting end. After the pre-compensated optical signal enters the single-fiber long-distance transmission, the pre-injected reverse phase component it carries will cancel out the approximately 0.12 rad positive parasitic phase shift caused by the fiber Kerr effect segment by segment. This reduces the residual nonlinear phase error of the first channel to near zero when it reaches the receiving end. The bit error rate tolerance limit for trip commands and differential protection sampling messages is as low as... The transmission reliability of high-priority data streams of this magnitude is thus improved. At the same time, since the pre-distortion at the sending end has eliminated nonlinear interference at the source, the residual deviation that needs to be corrected for the compensation vector in the closed-loop return of step five is reduced, and the iterative convergence speed is accelerated.

[0033] In step S23, based on the pre-distorted electrical signal, a tunable laser is driven to perform laser external modulation and photonic domain mapping to generate a set of pre-compensated optical signals. It should be noted that since the pre-distorted electrical signal generated in the aforementioned sub-step is still in the electrical domain and has not yet been converted into an optical signal that can be transmitted in an optical fiber, it is necessary to use laser external modulation to completely map the voltage waveform carrying the pre-compensation characteristics to the photonic domain in order to form independent wavelength optical signals that can be subsequently combined and transmitted through a single fiber. Based on this, the technical solution of this application further uses the pre-distorted electrical signal to drive a tunable laser to perform laser external modulation and photonic domain mapping to generate a set of pre-compensated optical signals. Through the above processing, the electrical signal with superimposed power gain and phase pre-compensation characteristics in the electrical domain can be faithfully mapped into independent optical signals for each channel, completing the physical conversion from the electrical domain to the photonic domain.

[0034] More specifically, in a specific example of this application, the tunable laser corresponding to each channel outputs continuous narrow-linewidth optical carriers according to the center wavelength standard value specified in the modified driving parameters. The center wavelength of each optical carrier is strictly locked to the pre-allocated frequency position in the ITU-T standard wavelength grid. The analog voltage waveforms of each channel in the pre-distorted electrical signal output from the aforementioned sub-step are fed into the Mach-Zehnder modulator of the corresponding channel. The modulator performs external modulation operation on the continuous optical carriers according to the instantaneous voltage amplitude and phase changes of the pre-distorted electrical signal, completely mapping the pre-distortion characteristics carried in the electrical domain, such as power gain, self-channel linear phase compensation, and cross-channel nonlinear interference pre-cancellation, to the amplitude and phase dimensions of the photonic domain. After modulation, each channel generates an optical signal with an independent center wavelength and pre-superimposed compensation. The modulated output optical beams of all channels are spatially combined and encapsulated to generate a pre-compensated optical signal set, which is then transferred to the next multiplexing stage.

[0035] Specifically, in step S3, the pre-compensated optical signal set undergoes multi-wavelength multiplexing processing at the pure optical physical layer, and the multiplexing result is coupled and superimposed with the out-of-band pilot signal independently generated in the non-service band to generate a telemetry-carrying composite optical flow for bidirectional transmission over a single fiber. It should be noted that since the pre-compensated optical signal set output in step two still exists as multiple independent wavelength beams and has not yet been integrated into a single-core fiber, and subsequent step four requires independent physical state monitoring of each wavelength channel at the receiving end, if the composite optical flow does not carry a telemetry identifier that can serve as a baseline reference, the receiving end will lack a time anchor point and amplitude reference for quantifying the degradation degree of a single channel. Based on this, the technical solution of this application further performs multi-wavelength multiplexing processing on the pre-compensated optical signal set at the pure optical physical layer, and couples and superimposes the multiplexing result with the out-of-band pilot signal independently generated in the non-service band to generate a telemetry-carrying composite optical flow for bidirectional transmission over a single fiber. Through the above processing, multiple pre-compensated optical signals and telemetry pilot signals can be integrated into a single-core optical fiber for transmission, providing an embedded physical reference for single-channel state feature extraction at the receiving end.

[0036] More specifically, in a particular example of this application, the pre-compensated optical signal set is first subjected to frequency domain channel precision alignment and pure optical physical layer multi-wavelength multiplexing processing based on standard wavelength grating specifications to generate a baseline composite optical flow. Then, an independent telemetry laser is driven to generate out-of-band pilot signals in non-service bands. Finally, the two are coupled and superimposed through a broadband wavelength division coupler to obtain a telemetry-carrying composite optical flow. In the frequency domain channel precision alignment and multi-wavelength multiplexing stage, the independent wavelength beams from the pre-compensated optical signal set output in step two are introduced into the arrayed waveguide grating of the dense wavelength division multiplexer. The center wavelengths of each optical signal are precisely aligned in the frequency domain according to the ITU-T standard wavelength grating specifications, ensuring that the wavelength spacing between adjacent channels strictly meets the grating spacing requirement of 0.8 nm or 0.4 nm. After alignment, the arrayed waveguide grating couples and constrains the pre-compensated optical signals with different center wavelengths to a single fiber channel at the pure optical physical layer. The optical power spectral densities of each wavelength optical signal are linearly superimposed in the frequency domain to generate a baseline composite optical flow carrying all service channels.

[0037] After the baseline composite optical flow is generated, the independent generation stage of the out-of-band pilot signal begins. An independent telemetry laser is driven to excite a continuous base optical carrier within a 1625 nm optical monitoring window, far from the core service data band. This base optical carrier is then mapped using a pseudo-random binary sequence with low-frequency amplitude modulation, enabling the pilot signal to carry the digital identity of the transmitting node and reference timestamp information, thus generating the out-of-band pilot signal. The center wavelength of this pilot signal maintains sufficient spectral isolation from the service band to avoid co-channel interference with the service channel.

[0038] After the out-of-band pilot signal is ready, the coupling and superposition stage begins. Using a broadband wavelength division coupler, the injected power of the out-of-band pilot signal is first attenuated and adjusted based on a secure optical power limit budget algorithm. This ensures that the pilot signal energy is sufficient to support cross-site telemetry demodulation while not exceeding the stimulated Raman scattering threshold, thus avoiding nonlinear crosstalk to the backbone service band. After attenuation adjustment, the power-controlled out-of-band pilot signal and the baseline composite optical flow are physically superimposed at the coupler's output port to generate a telemetry-carrying composite optical flow that simultaneously carries all service channels and telemetry pilots. This flow is then pushed into a dual-redundant fiber optic physical link to complete the transfer and delivery to step four.

[0039] Specifically, in step S4, the telemetry-carrying composite optical flow is transmitted using a dual-redundant all-optical link, and at the receiving end, a bypass splitter is used to non-intrusively extract the sampled optical flow for single-channel state feature extraction to obtain the arrival-state composite optical flow and fiber state matrix. It should be noted that, given that the telemetry-carrying composite optical flow output in step three needs to travel tens of kilometers of long-distance fiber optic transmission to reach the receiving end, sudden events such as fiber breakage or excessive bending during transmission will directly lead to communication interruption. Furthermore, the differentiated physical degradation information generated by dispersion accumulation and nonlinear effects in each wavelength channel during long-distance transmission is the data source for generating the compensation vector in step five. If the receiving end lacks the ability to finely perceive the physical state at the single-channel level, the entire closed-loop pre-compensation mechanism will lose its feedback basis. Based on this, the technical solution of this application further performs dual-redundant all-optical link transmission on the telemetry-carrying composite optical flow, and at the receiving end, a bypass splitter is used to non-intrusively extract the sampled optical flow for single-channel state feature extraction to obtain the arrival-state composite optical flow and fiber state matrix. Through the above processing, while ensuring zero-interruption service switching in the event of sudden failures such as fiber breakage, the granularity of link monitoring can be reduced from the traditional total optical power level to the single wavelength level, providing channel-by-channel physical damage quantification data for the deviation calculation and compensation vector generation in step five.

[0040] Figure 5 This document describes a flowchart illustrating a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. It describes the all-optical link dual-redundancy transmission of the telemetry composite optical flow, and the non-intrusive extraction of single-channel state features from the sampled optical flow using a bypass splitter at the receiving end to obtain the arrival-state composite optical flow and fiber state matrix. Figure 5As shown, step S4 includes: S41, routing the telemetry composite optical flow to a primary / backup dual-redundant parallel optical fiber network for cross-site pure optical transmission, and at the receiving end, using an asymmetric bypass splitter to extract energy at a preset splitting ratio to obtain the arrival state composite optical flow and the sampling optical data flow; S42, performing broadband photoelectric conversion on the sampling optical data flow, and using fast Fourier transform and digital bandpass filter bank for time-frequency domain separation and decoupling to obtain the pilot electrical sequence and the envelope characteristics of each wavelength; S43, based on the baseline time and amplitude anchor points provided by the pilot electrical sequence, performing differential comparison and physical damage quantization on the envelope characteristics of each wavelength to obtain the fiber state matrix.

[0041] In step S41, the telemetry-carrying composite optical flow is routed to a dual-redundant parallel optical fiber network for cross-station pure optical transmission. At the receiving end, an asymmetric bypass splitter is used to extract energy at a preset splitting ratio to obtain the arrival-state composite optical flow and the sampling optical data flow. It should be noted that since the transmission distance of the telemetry-carrying composite optical flow between substations is often tens of kilometers, there is a risk of sudden failures such as fiber breakage or excessive bending during transmission. Furthermore, the subsequent single-channel state feature extraction requires obtaining sampling optical energy from the transmission link without affecting the integrity of the backbone service signal. Based on this, the technical solution of this application further routes the telemetry-carrying composite optical flow to a dual-redundant parallel optical fiber network for cross-station pure optical transmission, and at the receiving end, an asymmetric bypass splitter is used to extract energy at a preset splitting ratio to obtain the arrival-state composite optical flow and the sampling optical data flow. Through the above processing, it is possible to obtain sampling optical energy for single-channel state monitoring in a non-intrusive manner while ensuring zero-interruption service switching under fiber optic fault conditions.

[0042] More specifically, in a specific example of this application, the optical protection board receives the telemetry-carrying composite optical flow output from step three and, based on optical line protection technology, distributes it according to a 1+1 primary and backup dual-transmission protection mechanism, simultaneously loading it onto the primary fiber optic link and the backup fiber optic link for parallel physical transmission. During transmission, the optical power, signal attenuation, and bit error rate of the primary and backup fiber optic links are continuously monitored in real time for 24 hours. If the primary link is normal, the signal arriving at the receiving end from the primary link is directly extracted and continues to flow downstream. If a fault is detected in the primary link, an automatic switching mechanism is triggered within tens of milliseconds to seamlessly intercept and extract the signal from the backup fiber optic link to maintain communication continuity. Upon arrival at the receiving end, the pre-positioned asymmetric bypass splitter intercepts the optical energy in the transmission link according to a preset splitting ratio of 99:1. 99% of the core backbone optical energy is directly transmitted through and defined as the arrival-state composite optical flow, and transferred to the pure optical physical demultiplexer in step five. The remaining 1% of weak optical energy is defined as the sampling optical data flow and directed to the subsequent time-frequency domain separation and decoupling stage. The received optical power of the sampled optical data stream is equal to the product of the bypass coupling coefficient and the total transmitted power at the origin, multiplied by the link attenuation factor determined by the fiber linear attenuation constant and the transmission distance. This power level must meet the minimum sensitivity requirements of the subsequent photoelectric conversion. At the same time, the extremely low interception ratio of 1% ensures that the energy loss of the composite optical stream arriving at the backbone is controlled within 0.04 dB, without having a substantial impact on the demodulation margin of the service signal.

[0043] In step S42, the sampled optical data stream undergoes broadband photoelectric conversion, and time-frequency domain separation and decoupling are performed using Fast Fourier Transform and a digital bandpass filter bank to obtain the pilot electrical sequence and envelope characteristics for each wavelength. It should be noted that since the sampled optical data stream simultaneously contains out-of-band pilot signals located in the 1625 nm non-service band and envelope information from various service wavelength channels, these two signals coexist in the same beam with different center wavelengths in the optical domain. Therefore, they need to be converted from the optical domain to the electrical domain and separated in the frequency domain to obtain the pilot sequence used for baseline anchoring and the envelope characteristics used for channel-by-channel degradation quantization, respectively. Based on this, the technical solution of this application further performs broadband photoelectric conversion on the sampled optical data stream and performs time-frequency domain separation and decoupling using Fast Fourier Transform and a digital bandpass filter bank to obtain the pilot electrical sequence and envelope characteristics for each wavelength. Through the above processing, telemetry reference information and physical state information of each service channel can be independently extracted from the aliased sampled light energy, providing separated input data for subsequent differential comparison and physical damage quantification.

[0044] More specifically, in a concrete example of this application, the sampled optical data stream is first subjected to broadband photoelectric conversion to obtain a hybrid analog electrical signal. Then, digital processing and frequency domain filtering are used to separate and decouple the pilot sequence from the envelope characteristics of each wavelength. During the broadband photoelectric conversion stage, the sampled optical data stream intercepted and diverted in the previous stage is introduced into a high-sensitivity avalanche photodiode. Utilizing the photoelectric effect, the sampled optical data stream is linearly converted into a hybrid analog photocurrent signal in the broadband physical domain. This photocurrent signal simultaneously contains the low-frequency modulation component corresponding to the pilot signal and the envelope fluctuation components corresponding to each service wavelength channel. Subsequently, the hybrid analog photocurrent signal undergoes high-speed analog-to-digital conversion, performing oversampling digital discretization processing at a rate higher than the minimum rate required by the Nyquist sampling theorem to generate a digitized hybrid electrical signal sequence.

[0045] After the digitized hybrid electrical signal sequence is ready, the time-frequency domain separation and decoupling stage begins. A Fast Fourier Transform (FFT) is performed on the digitized hybrid electrical signal sequence to map it from the time domain to the frequency domain, obtaining the complete spectral distribution. In the frequency domain, the out-of-band pilot signal, due to low-frequency amplitude modulation mapping, is concentrated in the low-frequency baseband region, exhibiting natural distributional isolation from the envelope spectral components of each service wavelength channel on the frequency axis. A digital bandpass filter bank is used to precisely segment the spectral distribution, extracting the spectral components in the low-frequency baseband region that match the 1625 nm pilot modulation frequency and performing an inverse transform to restore the time-domain pilot electrical sequence. This sequence carries the reference timestamp and amplitude reference information injected by the transmitter. Simultaneously, the digital bandpass filter bank is used to lock the spectral components at the frequency points of the center wavelengths of each backbone service, independently extracting the level envelope fluctuations of each service channel, decoupling the amplitude attenuation characteristics and phase shift jitter characteristics of each channel, and generating the envelope features of each wavelength. The pilot electrical sequence and the envelope features of each wavelength are output in parallel to the subsequent physical damage quantization stage.

[0046] In step S43, based on the baseline time and amplitude anchor points provided by the pilot electrical sequence, differential comparison and physical impairment quantization are performed on the envelope features of each wavelength to obtain the fiber state matrix. It should be noted that since the pilot electrical sequence and the envelope features of each wavelength output in the previous step respectively carry the baseline reference information injected by the transmitter and the actual physical state information of each channel after long-distance transmission, and these two have not yet been correlated and compared, they cannot be directly used for the calculation of the compensation vector. The pilot sequence needs to be used as a unified reference anchor point to quantify the physical degradation degree of each channel. Based on this, the technical solution of this application further performs differential comparison and physical impairment quantization on the envelope features of each wavelength based on the baseline time and amplitude anchor points provided by the pilot electrical sequence to obtain the fiber state matrix. Through the above processing, the deviation between the actual received state of each channel and the transmitter reference can be quantized into structured matrix data, providing channel-by-channel physical impairment input for the deviation calculation in step five.

[0047] More specifically, in a concrete example of this application, the reference timestamp and reference amplitude reference value injected by the transmitter are first parsed from the pilot electrical sequence, serving as a globally unified time alignment anchor and amplitude comparison baseline. Subsequently, the actual received envelope amplitude of each independent service channel in the envelope characteristics of each wavelength is read, and differentially compared with the amplitude of the expected reference baseline mapped by the pilot baseline anchor point for that wavelength, quantifying and calculating the independent microphysical attenuation coefficient of each single-wavelength channel. The microphysical attenuation coefficient of the k-th single-wavelength channel is then calculated. It equals a constant -10 multiplied by a logarithm to base 10, where the logarithm represents the actual electrical envelope amplitude extracted for that wavelength. Divide by the amplitude of the reference baseline at that wavelength obtained by mapping the pilot sequence. In addition, the global macroscopic attenuation of the pilot channel itself ,Right now: in, To extract the actual electrical envelope detector amplitude of the decoupled k-th wavelength channel from the envelope characteristics of each wavelength, in V. The expected reference baseline amplitude for the k-th wavelength channel, calculated based on the timestamp characteristics of the pilot electrical sequence and a preset transmission baseline mapping, is expressed in V. The global macroscopic basis attenuation of the fiber optic link, measured in dB, is directly calculated from the amplitude drop of the pilot sequence and is used to correct for the overall translational deviation caused by relative measurements. After quantizing the attenuation coefficients of each channel, the delay drift of the envelope characteristics of each wavelength relative to the pilot timing is calculated simultaneously and converted into phase shift jitter values ​​for each wavelength to reflect the degree of dispersion accumulation. The independent microscopic physical attenuation coefficients and phase shift jitter values ​​covering all effective channels are tensor-encapsulated according to a two-dimensional structure of wavelength index and physical impairment type. The row index corresponds to the wavelength number of each channel, and the column index corresponds to the two types of physical quantities: attenuation coefficient and phase shift jitter value. This generates the fiber state matrix and is transferred to step five for use in the multinomial regression model calculation.

[0048] Specifically, in step S5, the arrival-state composite optical flow is demodulated to output the service signal. Simultaneously, single-wavelength deviation calculation is performed on the fiber state matrix to generate a compensation vector, which is then loop-backed to step two via the reverse optical path. It should be noted that, given that step four has separated the telemetry-carrying composite optical flow into the arrival-state composite optical flow carrying all service channels and the fiber state matrix quantified with channel-by-channel physical impairments, the service data of each wavelength channel in the arrival-state composite optical flow needs to be demodulated to be restored into independent service signals that can be directly used by the terminal equipment. The attenuation coefficients and phase shift jitter values ​​of each channel recorded in the fiber state matrix need to be calculated and converted into executable correction instructions, which are then loop-backed to the transmitter via the reverse optical path to drive step two to perform targeted power and phase pre-distortion adjustments for each channel in the next communication cycle, thereby closing the entire out-of-band telemetry and dynamic pre-compensation feedback loop. Based on this, the technical solution of this application further performs wavelength division demodulation on the composite optical flow of arrival to output the service signal, and simultaneously performs single-wavelength deviation calculation on the fiber state matrix to generate a compensation vector, which is then looped back to step two via the reverse optical path. Through the above processing, while the lossless restoration of the service signal and terminal delivery can be completed at the pure optical layer, the channel-by-channel physical degradation state perceived by the receiver can be continuously transformed into pre-compensation commands that can be directly used by the transmitter, enabling the entire transmission link to form an adaptive closed-loop repair capability between the transmitting and receiving ends without interrupting the service.

[0049] Figure 6 This document describes a flowchart of a single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to an embodiment of this application. The method involves WDM demodulating the arriving composite optical flow to output a service signal, simultaneously performing single-wavelength deviation calculations on the fiber state matrix to generate a compensation vector, and then looping the compensation vector back to step two via a reverse optical path. Figure 6 As shown, step S5 includes: S51, based on the inherent diffraction deflection angle differences of each wavelength, performing lossless wavelength division demodulation and direct output of the arrival state composite optical flow at the pure optical physical layer and the optical layer interface to obtain the service signal; S52, performing nonlinear deviation calculation on the physical attenuation and phase shift jitter values ​​of each channel in the fiber state matrix to generate a compensation vector; S53, through the optical monitoring channel transmitter, performing forward error correction coding and reverse optical path modulation back on the compensation vector, so as to send the compensation vector to the transmitting end through the reverse communication link closed loop for use in step two.

[0050] In step S51, based on the inherent diffraction deflection angle differences of each wavelength, the arriving state composite optical flow undergoes lossless wavelength division demodulation at the pure optical physical layer and direct output through the optical layer interface to obtain the service signal. It should be noted that since the arriving state composite optical flow still carries various service data such as tripping command streams, high-frequency sampling messages, and video frame sequences in a multi-wavelength aliasing manner, the terminal protection device cannot directly identify and extract the service content at the corresponding wavelength from the composite optical flow. It is necessary to strip each wavelength channel back into independent optical signals at the pure optical layer before it can be used. Based on this, the technical solution of this application further utilizes the inherent diffraction deflection angle differences of each wavelength to perform lossless wavelength division demodulation at the pure optical physical layer and direct output through the optical layer interface to obtain the service signal. Through the above processing, the service data of each channel can be restored into independent optical signals that the terminal can directly use without photoelectric-optical conversion, avoiding signal distortion and additional losses introduced by electrical layer processing.

[0051] More specifically, in a specific example of this application, the arriving state composite optical flow transmitted in step four is introduced into the arrayed waveguide grating demultiplexer within the receiver's multiplexing / demultiplexing board. The arriving state composite optical flow propagates within the free-space diffraction region. Due to their different center wavelengths, each wavelength optical signal has an inherent difference in diffraction deflection angle. The deflection angles of shorter wavelength optical signals and longer wavelength optical signals are naturally separated in space. Thus, the aliased composite optical flow is losslessly stripped into multiple independent single-wavelength beams at the pure optical physical layer. Each single-wavelength beam converges to the output waveguide port corresponding to the arrayed waveguide grating and is directly output through a standard optical layer interface as a service signal that can be directly photoelectrically resolved by each terminal device. The entire wavelength division and demodulation process is completed at the pure optical physical layer, eliminating the need to convert the optical signal to an electrical signal and then remodulate it back to an optical signal. This avoids quantization noise and time delay jitter during photoelectric-optical conversion. Furthermore, it eliminates the need to replace boards or optical modules on the service side. Each service signal can be directly delivered to substation protection devices, video monitoring terminals, and dispatch telephone equipment after being directly output through the optical layer interface.

[0052] In step S52, a nonlinear deviation calculation is performed on the physical attenuation and phase shift jitter values ​​of each channel in the fiber state matrix to generate a compensation vector. It should be noted that since the microscopic physical attenuation coefficients and phase shift jitter values ​​of each channel recorded in the fiber state matrix are only raw physical observations and have not yet been converted into correction instructions that can directly drive the transmitting end optical wavelength conversion board to perform power and wavelength fine-tuning, and there is a nonlinear coupling relationship between attenuation and phase shift in the fiber link, simple linear proportional conversion cannot accurately reflect the actual compensation requirements. Based on this, the technical solution of this application further performs a nonlinear deviation calculation on the physical attenuation and phase shift jitter values ​​of each channel in the fiber state matrix to generate a compensation vector. Through the above processing, the raw physical observations can be converted into a structured compensation vector containing a channel-by-channel power compensation component, providing accurate correction instructions for the pre-distortion conversion in the next communication cycle in step two.

[0053] More specifically, in a concrete example of this application, the microphysical attenuation coefficient and phase shift jitter value of each effective channel in the fiber state matrix are read, and a pre-fitted polynomial regression model based on fiber degradation samples is invoked to perform nonlinear deviation calculations on the power compensation components of each channel. The power compensation component of the k-th wavelength channel... Equal to static foundation compensation constant Add first-order attenuation coefficient Compared with the actual physical attenuation of the channel The product of the two, plus the second-order nonlinear decay coefficient Square of the actual physical attenuation of the channel The product of these factors, plus the dispersion compensation coefficient. Phase shift jitter value of the channel The product of, i.e.: in, This is the static base compensation constant used to offset the static losses of the link base, in dB. Here, represents the first-order linear decay coefficient fitted to the regression model, and represents the dimensionless weights. This represents the actual physical attenuation of the k-th channel extracted from the fiber state matrix, in dB. Here, is the second-order nonlinear attenuation coefficient used to compensate for the nonlinear effects of large signals in optical fibers, and is a dimensionless weight. To convert the phase shift into a dispersion compensation coefficient for power compensation redundancy, This represents the phase shift jitter value of the k-th channel extracted from the fiber state matrix, in rad. In this formula, the first-order term compensates for the linear power drop proportional to the attenuation, the second-order term captures the additional loss caused by stimulated Raman scattering and other effects in long-distance transmission that is non-linearly related to the attenuation, and the dispersion term converts the pulse broadening effect caused by dispersion accumulation into an equivalent power margin requirement. After completing the polynomial regression calculation for all effective channels, the power compensation components of each channel are encapsulated into a structured compensation vector according to the channel index, and then transferred to the subsequent reverse optical path closed-loop backhaul stage.

[0054] Specifically, the coefficients in the above polynomial regression model , , and The settings are not based on human experience, but rather obtained through offline fitting training during the equipment's factory calibration phase. During calibration, the optical transmission equipment to be deployed is connected to standard test fibers of different lengths and models. Under controlled experimental conditions, variables such as ambient temperature, injected optical power, and the number of channels are gradually adjusted to subject the fiber optic link to various physical states, ranging from slight attenuation to severe degradation. Under each set of variables, the actual microscopic physical attenuation coefficient and phase shift jitter value of each channel are collected at the receiver using a high-precision spectral analyzer. Simultaneously, the optimal power compensation amount required to restore the receiver's bit error rate to below the target threshold under that degradation state is recorded. After collecting data under hundreds to thousands of different degradation conditions, the microscopic physical attenuation coefficient and phase shift jitter value of each channel are used as independent variables, and the corresponding optimal power compensation amount is used as the dependent variable. The least squares method is used to fit the parameters of the polynomial model, solving for the coefficient combination that minimizes the sum of squared residuals between the predicted compensation amount and the actual optimal compensation amount at all sample points, thus determining the optimal power compensation amount. , , and The values ​​are then fixed and written to the read-only storage area of ​​the receiver's diagnostic algorithm engine after fitting. This allows for direct access during actual device operation without requiring repeated training.

[0055] In step S53, the compensation vector is forward-corrected and modulated back via an optical monitoring channel transmitter to deliver it to the transmitting end via a closed-loop reverse communication link for use in step two. It should be noted that since the compensation vector generated in the previous step is located at the receiving end, while step two, which requires consuming this vector, is located at a remote transmitting node, the compensation vector must be transmitted back via a reverse optical path to close the entire pre-compensation feedback loop. Furthermore, the reverse transmission link also faces noise interference and bit error risks. If bit flipping occurs during the transmission of the compensation vector, the transmitting end will receive an incorrect correction command. Therefore, the technical solution of this application further utilizes an optical monitoring channel transmitter to perform forward-correction coding and reverse optical modulation back on the compensation vector, delivering it to the transmitting end via a closed-loop reverse communication link for use in step two. Through the above processing, the data integrity of the compensation vector during reverse transmission can be ensured, enabling the transmitting end to receive a lossless correction command in the next communication cycle, thereby closing the adaptive pre-compensation feedback loop between the transmitting and receiving ends.

[0056] More specifically, in a concrete example of this application, the power compensation components of each channel in the compensation vector are first subjected to serial binary quantization, converting the floating-point precision numerical sequence into a fixed-point binary bit stream. Then, forward error correction coding is performed on this bit stream, embedding redundant parity bits to resist sudden noise interference during reverse transmission. The encoded bit stream has the ability to autonomously detect and correct a certain number of transmission errors at the receiving end. After coding, the local optical monitoring channel transmitter at the receiving end is driven to modulate the encoded bit stream carrying compensation information onto an independent reverse monitoring optical wave. The center wavelength of this monitoring optical wave is selected at the 1510 nm reverse monitoring channel, spectrally isolated from the forward service transmission band. The modulated reverse monitoring optical signal is transmitted to the originating node at the other end through a single-fiber bidirectional reverse communication link in a passive optical network. After receiving the reverse monitoring optical signal, the other node performs photoelectric conversion and forward error correction decoding to restore the lossless compensation vector, which is then written into a local buffer register as the input parameter for step two of the next service cycle. This completes the end-to-end bidirectional closed loop of physical transmission and data compensation. In the first operating cycle, since there is no feedback data from the receiver, the compensation vector is initialized as an all-zero matrix. As the communication cycle continues to iterate, the power compensation components of each channel in the compensation vector gradually approach the steady-state value that matches the current link degradation state.

[0057] As described above, the single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks according to embodiments of this application can be implemented in various power communication optical transmission and intelligent power grid operation and maintenance scenarios, such as fiber optic communication equipment rooms between substations, optical transmission aggregation nodes in distribution networks, or optical relay stations along transmission lines. In one possible implementation, this method can be integrated into a power communication network management platform or optical transmission network management information system as a single-fiber WDM optical transmission and closed-loop adaptive compensation engine. For example, this method can be an independent pre-distortion calculation and link monitoring application running on the embedded controller of the optical transmission equipment in the communication room, or it can be an out-of-band telemetry sensing and reverse compensation closed-loop processing component of the existing wavelength division multiplexing optical transmission equipment, or it can be an intelligent optical network operation and maintenance middleware service deployed on the dispatch master station side and issuing compensation commands through the optical monitoring channel. Of course, the protocol splitting parsing and baseline parameter extraction, wavelength adaptive pre-distortion conversion, and single wavelength deviation calculation components based on polynomial regression in this method can also run on the optical wavelength conversion board, multiplexing and splitting board, or optical protection board deployed in the communication room of the substation, as the underlying real-time optical link status sensing and compensation execution base of this single-fiber bidirectional wavelength division transmission system.

[0058] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A single-fiber bidirectional wavelength division multiplexing (WDM) transmission method for passive optical networks, characterized in that, include: Step 1: Perform protocol splitting and parsing and service level baseline parameter extraction on the multi-protocol data stream collected from the access side, which includes tripping command stream, high-frequency sampling message and video frame sequence, to obtain the service load matrix and driving parameter set; Step 2: Based on the driving parameter set and the compensation vector returned by the closed loop at the receiver, perform wavelength adaptive pre-distortion transformation on the service payload matrix to obtain an independent set of pre-compensated optical signals for each channel. Step 3: Perform multi-wavelength multiplexing on the pre-compensated optical signal set in the pure optical physical layer, and couple and superimpose the multiplexing result with the out-of-band pilot signal generated independently in the non-service band to generate a telemetry composite optical flow for single-fiber bidirectional transmission. Step 4: Perform all-optical link dual-redundancy transmission on the telemetry composite optical flow, and at the receiving end, use a bypass splitter to non-intrusively intercept the sampled optical flow for single-channel state feature extraction to obtain the arrival state composite optical flow and fiber state matrix. Step 5: Demodulate the arrival state composite optical flow to output the service signal, and at the same time perform single-wavelength deviation calculation on the fiber state matrix to generate a compensation vector, and then transmit the compensation vector back to Step 2 via the reverse optical path closed loop.

2. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 1, characterized in that, Step one includes: Step 1.1: Based on the Ethernet type field and VLAN tag characteristics, perform protocol slicing and priority field parsing on the multi-protocol data stream to obtain the classified message sequence and service level identifier; Step 1.2: Perform multi-layer protocol header stripping on the classified message sequence to extract the pure data segment, and perform two-dimensional reassembly mapping on the pure data segment according to the time slice and spatial channel dimensions to obtain the service payload matrix; Step 1.3: Based on the priority weight in the service level identifier and the system reference minimum optical power, perform reference transmit power calculation and bind the center wavelength standard value to each channel to generate a driving parameter set.

3. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 1, characterized in that, Step two includes: Step 2.1: Use the driving parameter set as a reference tensor and perform channel-by-channel tensor superposition with the power compensation components of each channel in the compensation vector to obtain the corrected driving parameters. Step 2.2: Based on the corrected driving parameters, the service load matrix is ​​subjected to electrical domain pre-distortion waveform reconstruction and digital-to-analog conversion processing to generate a pre-distortion electrical signal; Step 2.3: Based on the pre-distorted electrical signal, drive the tunable laser to perform laser external modulation and photonic domain mapping to generate a pre-compensated optical signal set.

4. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 1, characterized in that, Step three includes: Step 3.1: Based on the standard wavelength grid specification, perform frequency domain channel precision alignment and pure optical physical layer multi-wavelength multiplexing on the pre-compensated optical signal set to generate baseline composite optical flow; Step 3.2: Drive an independent telemetry laser to excite the substrate optical carrier in the non-service band, and use a pseudo-random binary sequence to perform low-frequency amplitude modulation mapping on the substrate optical carrier to obtain an out-of-band pilot signal. Step 3.3: Using a broadband wavelength division coupler, the out-of-band pilot signal is injected and attenuated based on the safe optical power limit budget, and coupled and superimposed with the baseline composite optical flow to obtain the telemetry composite optical flow.

5. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 1, characterized in that, Step four includes: Step 4.1: The telemetry composite optical flow is routed to the primary and backup dual-redundant parallel optical fiber network for cross-site pure optical transmission, and at the receiving end, an asymmetric bypass splitter is used to extract energy according to a preset splitting ratio to obtain the arrival state composite optical flow and the sampling optical data flow. Step 4.2: Perform broadband photoelectric conversion on the sampled optical data stream, and perform time-frequency domain separation and decoupling through fast Fourier transform and digital bandpass filter bank to obtain the pilot electrical sequence and the envelope characteristics of each wavelength. Step 4.3: Based on the baseline time and amplitude anchor points provided by the pilot electrical sequence, perform differential comparison and physical damage quantization on the envelope features of each wavelength to obtain the fiber state matrix.

6. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 1, characterized in that, Step five includes: Step 5.1: Based on the inherent diffraction deflection angle differences of each wavelength, the arrival state composite optical flow is subjected to lossless wavelength division demodulation by the pure optical physical layer and direct output through the optical layer interface to obtain the service signal. Step 5.2: Perform nonlinear deviation calculation on the physical attenuation and phase shift jitter values ​​of each channel in the fiber state matrix to generate a compensation vector; Step 5.3: The compensation vector is forward error correction coding and reverse optical path modulation back through the optical monitoring channel transmitter, so as to send the compensation vector to the transmitting end through the reverse communication link closed loop for use in step two.

7. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 3, characterized in that, Step 2.2 includes: Based on the corrected drive parameters, determine the voltage gain coefficient and the phase compensation bias. Based on the voltage gain coefficient and phase compensation bias, the service load matrix is ​​subjected to electrical domain pre-distortion waveform reconstruction and digital-to-analog conversion to generate a pre-distortion electrical signal.

8. The single-fiber bidirectional wavelength division transmission method for passive optical networks according to claim 3, characterized in that, Step 2.2 includes: The channel coupling weight matrix is ​​determined based on the modified driving parameters; Based on the channel coupling weight matrix, cross-channel nonlinear interference phase prediction is performed on the service payload matrix to obtain the cross-channel interference phase vector; Based on the cross-channel interference phase vector and the corrected driving parameters, the traffic payload matrix is ​​coupled to sense and reconstruct the pre-distorted waveform to obtain the pre-distorted electrical signal.

Images