A method, device and medium for processing optical signals in a passive optical network

Abstract

The application discloses a kind of optical signal processing method, equipment and medium in passive optical network, it is related to passive optical network signal processing technical field, comprising: the digital sampling signal of last uplink burst end, guard interval and current uplink burst preamble section is collected, removes reconfigurable communication signal, obtains observation sequence and the original sampling cache of current uplink burst;Select candidate time constant, establish exponential base function matrix, and carry out detrending and amplitude normalization;Perform sparse estimation with amplitude boundary and initial continuity constraint, obtain candidate memory component;Merge candidate memory component, screen effective memory component;Predict the residual in current uplink burst, determine target compensation ratio using multiple candidate compensation ratios;According to target compensation ratio, process uplink burst, when verification fails, call original sampling cache to execute uncompensated processing.The application realizes compound recovery tail representation and adaptive determination of compensation degree, guarantees reliable reception of uplink burst optical signal.

Description

Technical Field

[0001] This invention relates to the field of passive optical network signal processing technology, and in particular to a method, device and medium for optical signal processing in a passive optical network. Background Technology

[0002] Passive optical networks (PONs) employ a point-to-multipoint fiber access architecture, with the optical line terminal (OLT) centrally managing the uplink transmissions of multiple optical network units. Due to variations in transmission distance, transmit power, and device status, the uplink typically receives signals in burst mode. Conventional optical signal processing methods generally perform burst detection based on licensed time slots and preambles, followed by sequential gain adjustment, clock recovery, adaptive equalization, and decoding. For baseline changes between adjacent bursts or the recovery process at the receiver front end, processing is performed using guard intervals, fixed recovery parameters, or preset compensation amounts.

[0003] The above-mentioned conventional methods have two main problems: First, different burst power differences, guard intervals and gain states may form recovery wakes superimposed on multiple time scales, and fixed parameters or a single time constant are difficult to characterize the actual residual response; Second, the preset compensation amount usually lacks a joint verification and back-off mechanism based on preamble quality, equalization error and decoding results. When the estimation deviation is large, it is easy to cause insufficient compensation or overcompensation, which affects the reliable reception of the current uplink burst. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention is proposed.

[0005] Therefore, this invention provides an optical signal processing method in a passive optical network, which solves the problems of insufficient characterization of composite recovery wakes and lack of joint verification and backoff mechanism for compensation results in the prior art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an optical signal processing method in a passive optical network, comprising: acquiring digital sampled signals of the end of the previous uplink burst, the guard interval, and the preamble of the current uplink burst; removing reconfigurable communication signals to obtain the observation sequence and the original sampling buffer of the current uplink burst; selecting candidate time constants based on receiver calibration results and the operating status of adjacent uplink bursts; establishing an exponential basis function matrix; and performing detrending and amplitude normalization on the observation sequence; based on the exponential basis function matrix, performing sparse estimation with amplitude boundaries and initial continuity constraints on the normalized observation sequence to obtain candidate memory components; and according to the exponential basis function matrix... Basis function correlation merges candidate memory components with the same sign and closest similarity, and selects effective memory components based on residual contribution, cross-window stability, and fitting residual ratio. Based on the effective memory components, the residual amount in the current uplink burst is predicted, and multiple candidate compensation ratios are used to process the current uplink burst preamble. The target compensation ratio is determined based on the preamble correlation quality, DC residual, and initial equalization error. The current uplink burst is processed according to the target compensation ratio, and gain adjustment, clock recovery, adaptive equalization, and decoding are performed in sequence. When the verification is successful, communication data is output and historical coefficients are updated. When the verification fails, the original sampling buffer is called to perform uncompensated processing.

[0008] As a preferred embodiment of the optical signal processing method in the passive optical network described in this invention, the specific steps for obtaining the observation sequence and the original sampling buffer of the current uplink burst are as follows: Based on the uplink grant time, the digital sampling amplitude change edge, and the preamble, determine the end position of the previous uplink burst, the guard interval, and the coarse start position of the current uplink burst; based on the decision symbols, amplitude estimation, equivalent received impulse response, and timing offset stored in the previous uplink burst, reconstruct the expected communication signal at the end of the previous uplink burst, and subtract the expected communication signal from the original sampling signal to obtain the end residual and the initial residual value; based on the known preamble sequence, coarse timing, and preliminary amplitude estimation of the current uplink burst, reconstruct the expected communication signal of the preamble segment of the current uplink burst, and subtract the expected communication signal from the original sampling signal of the preamble segment to obtain the preamble segment residual; arrange the guard interval sampling signal and the preamble segment residual in actual time order to form an observation sequence, and write the complete original sampling signal of the current uplink burst into the original sampling buffer.

[0009] As a preferred embodiment of the optical signal processing method in the passive optical network described in this invention, the specific steps of establishing the exponential basis function matrix and performing detrending and amplitude normalization on the observation sequence are as follows: First, determine the range of candidate time constants based on the receiver step response and the recovery curve after the test burst ends, and set candidate time constants at logarithmic intervals within the candidate time constant range. Second, select the candidate time constants to be used in the current processing cycle from the candidate time constants based on the power difference between preceding and following uplink bursts, the duration of the preceding uplink burst, the guard interval length, and the current receiver gain state. Third, establish the exponential basis function matrix based on the elapsed time of each observation sampling point in the observation sequence relative to the end time of the preceding uplink burst and the selected candidate time constants, and delete exponential basis functions that reach a preset degree with the constant baseline or linear baseline. Fourth, estimate the baseline constant term and baseline change rate of the observation sequence using robust linear regression, and perform detrending and amplitude normalization on the observation sequence based on the amplitude difference between the amplitude at the end of the preceding uplink burst and the amplitude of the guard interval baseline.

[0010] As a preferred embodiment of the optical signal processing method in the passive optical network described in this invention, the step of performing sparse estimation with amplitude boundaries and initial continuity constraints on the normalized observation sequence to obtain candidate memory components includes the following steps: Based on the difference in uplink burst power, guard interval length, and receiver gain status, historical coefficients stored under the same operating conditions are read, and these historical coefficients are used as the initial values ​​for sparse estimation iterations; an observation weight matrix is ​​set, where normal sampling points in the guard interval, residual sampling points in the current uplink burst preamble, and low-confidence sampling points correspond to different observation weights; the weighted fitting error between the candidate coefficient vector and the normalized observation sequence, and the norm of the candidate coefficient vector are used as optimization objects, and amplitude boundaries and initial continuity constraints corresponding to the normalized initial residual values ​​are set for each candidate coefficient; a fast iterative shrinkage threshold algorithm is used to solve for the candidate coefficients, and candidate coefficients with absolute values ​​lower than the noise amplitude threshold are set to zero, while the remaining candidate coefficients are combined with their corresponding candidate time constants to form candidate memory components.

[0011] As a preferred embodiment of the optical signal processing method in the passive optical network of the present invention, the step of merging candidate memory components with the same sign and closest correlation based on the correlation of exponential basis functions specifically includes: calculating the normalized correlation coefficient between the exponential basis functions of any two candidate memory components; when the normalized correlation coefficient between the two exponential basis functions reaches the component similarity threshold and the corresponding candidate coefficients have the same sign, the corresponding candidate memory components are grouped into the same group; for the same group, the candidate time constant corresponding to the candidate coefficient with the largest absolute value is determined as the equivalent time constant, and the amplitudes of each candidate coefficient in the group are added together and restored to the equivalent amplitude according to the amplitude normalization benchmark; when the normalized correlation coefficient between the two exponential basis functions reaches the component similarity threshold and the corresponding candidate coefficients have opposite signs, the corresponding candidate memory components are retained for subsequent screening.

[0012] As a preferred embodiment of the optical signal processing method in the passive optical network described in this invention, the specific steps for selecting effective memory components based on residual contribution, cross-window stability, and fitting residual ratio are as follows: Each candidate memory component is removed sequentially while keeping the remaining candidate memory components unchanged; the normalized residual energy before and after removal is calculated; the residual contribution of each candidate memory component is determined based on the change in normalized residual energy; the observation sequence is divided into at least two overlapping sub-windows; local sparsity estimation is performed within each sub-window based on candidate exponential basis functions; cross-window stability is determined based on the number of times a candidate memory component is retained in at least two sub-windows; the fitting residual ratio is calculated between the normalized residual energy using all candidate memory components and the normalized baseline residual energy not using candidate memory components; and candidate memory components whose residual contribution reaches a contribution threshold, cross-window stability reaches a stability threshold, fitting residual ratio does not exceed a fitting residual ratio threshold, and equivalent amplitude is not lower than a noise amplitude threshold are determined as effective memory components.

[0013] As a preferred embodiment of the optical signal processing method in the passive optical network described in this invention, the specific steps of determining the target compensation ratio based on the preamble correlation quality, DC residual, and initial equalization error are as follows: Based on the current uplink burst preamble and the original sampled signal, perform interpolation correlation search at integer sampling positions to obtain the fractional sampling level timing correction; predict the residual amount of each effective memory component at each sampling point of the current uplink burst based on the equivalent amplitude, equivalent time constant, guard interval time, and fractional sampling level timing correction of each effective memory component, and add the residual amounts at the same sampling points to obtain the total predicted residual amount; set at least two candidate compensation values ​​between zero and one. The proportions are calculated by subtracting the total predicted residual from the original sampled signal of the current uplink burst preamble according to each candidate compensation proportion. The normalized preamble correlation quality, normalized high-low level spacing, normalized DC residual, and normalized initial equalization error of each candidate compensation result are calculated respectively. Candidate compensation proportions that do not meet the preamble correlation improvement requirements, have high-low level spacing degradation exceeding the allowable range, or have normalized DC residual exceeding the allowable range are eliminated in sequence. Among the remaining candidate compensation proportions, the candidate compensation proportion with the smallest normalized initial equalization error is determined as the target compensation proportion. If no candidate compensation proportion meets the conditions, the proportion that does not perform compensation is determined as the target compensation proportion.

[0014] As a preferred embodiment of the optical signal processing method in the passive optical network of the present invention, the step of processing the current uplink burst according to the target compensation ratio and sequentially performing gain adjustment, clock recovery, adaptive equalization, and decoding is as follows: The complete original sampled signal of the current uplink burst is read from the original sampling buffer; the total prediction residual is deducted according to the target compensation ratio; and gain adjustment, clock recovery, adaptive equalization, and decoding are performed sequentially. The compensation processing result is verified based on whether the adaptive equalizer converges within a specified number of iterations, whether the equalization residual is within the residual range corresponding to the normally functioning sample of the receiver, and whether the forward error correction decoding or frame check passes. When the compensation processing result passes verification, the communication data after compensation processing is output; the current effective coefficients are formed according to the position of the effective memory component in the exponential basis function matrix, and the historical coefficients under the same operating conditions are updated using the forgetting coefficients. When the compensation processing result fails verification, the original sampling buffer is called to perform uncompensated processing without deducting the total prediction residual; when the uncompensated processing passes decoding verification, the communication data after uncompensated processing is output without updating the historical coefficients; when both compensation processing and uncompensated processing fail decoding verification, a current uplink burst reception failure flag is output.

[0015] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the optical signal processing method in a passive optical network as described in the first aspect of the present invention.

[0016] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step of the optical signal processing method in a passive optical network as described in the first aspect of the present invention.

[0017] The beneficial effects of this invention are as follows: by constructing a multi-timescale exponential basis function matrix, accurate characterization of the composite recovery wake is achieved; by constrained sparse estimation and effective component screening, effective separation of the recovery response from noise and baseline drift is achieved; by comparing candidate compensation ratios and constraining compensation quality, adaptive determination of the compensation degree is achieved; and by jointly verifying the compensation results and backing up the original sampling buffer, reliable reception of uplink burst optical signals is achieved. Attached Figure Description

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

[0019] Figure 1 This is a flowchart of an optical signal processing method in a passive optical network.

[0020] Figure 2 This is a flowchart illustrating the formation of the current uplink burst observation sequence.

[0021] Figure 3 A flowchart for efficient memory component estimation and filtering.

[0022] Figure 4 A flowchart for determining the target compensation ratio and verifying the rollback. Detailed Implementation

[0023] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0024] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0025] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0026] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides an optical signal processing method in a passive optical network, comprising the following steps:

[0027] S1. Collect the digital sampling signals of the end of the previous uplink burst, the protection interval, and the leading segment of the current uplink burst, remove the reconfigurable communication signals, and obtain the observation sequence and the original sampling buffer of the current uplink burst.

[0028] During the uplink authorized time slot, the optical line terminal uses an optical receiver to convert the uplink burst optical signal sent by the optical network unit into an electrical signal. After transimpedance amplification, gain adjustment and analog-to-digital conversion, a continuous digital sampling signal is obtained.

[0029] Furthermore, based on the uplink authorization time issued by the dynamic bandwidth allocation module, the nominal arrival time range of the current uplink burst is determined; within the nominal arrival time range of the current uplink burst, based on the digital sampling amplitude change edge and preamble correlation results, the end position of the previous uplink burst, the guard interval, and the coarse start position of the current uplink burst are determined.

[0030] Furthermore, the last sampling signal at the end position of the previous uplink burst, the guard interval sampling signal, and the leading segment sampling signal of the current uplink burst are extracted from the continuous digital sampling signal.

[0031] 1.1 Reconstruct the communication signal at the end of the previous uplink burst.

[0032] Read the decision symbols, amplitude estimation results, equivalent received impulse response, timing offset, and burst end position saved after the previous uplink burst has been received and processed.

[0033] Based on the data saved after the previous uplink burst was received and processed, the expected communication signal at the end of the previous uplink burst is reconstructed and represented as:

[0034] ;

[0035] in, Indicates the last of the previous upward sudden event. The reconstructed signal of each sampling point is expressed in units of analog-to-digital converter output code value. Indicates the sampling point number at the end of the previous uplink burst; This represents the amplitude estimate of the previous uplink burst, expressed in analog-to-digital converter output code values. This represents the set of tail symbol indices that participated in the reconstruction. Indicates the tail symbol sequence number; Indicates the first Each decision code element is a dimensionless numerical value; It represents the normalized equivalent received impulse response of the previous uplink burst, and is a dimensionless function; Indicates the sampling period, in seconds; Indicates the symbol period, in seconds; This indicates the timing offset of the previous uplink burst, in seconds.

[0036] It should be noted that the decision code uses the code after forward error correction decoding or frame check; when the previous uplink burst fails the decoding check, the high confidence decision code output by the equalizer is used.

[0037] Furthermore, a high-confidence decision symbol refers to a symbol whose distance from the decision point to the corresponding ideal level or ideal constellation point is less than the upper quantile value of the distance to the normal decoding sample, such as 95% of the symbols. When the number of high-confidence decision symbols is insufficient to cover the reconstruction interval at the end of the previous uplink burst, the residual at the end of the previous uplink burst is not used to establish the subsequent start continuity constraint.

[0038] The reconstructed signal is subtracted from the original sampled signal at the end of the previous uplink burst to obtain the terminal residual, which is expressed as:

[0039] ;

[0040] in, Indicates the last of the previous upward sudden event. The residual of each sampling point, in units of analog-to-digital converter output code value; Indicates the last of the previous upward sudden event. The original sampled values ​​of each sampling point are expressed in units of analog-to-digital converter output code values.

[0041] Based on the trailing residual at the end position of the previous uplink burst, the initial residual value of the leading composite wake is calculated and expressed as:

[0042] ;

[0043] in, This represents the initial residual value at the end of the previous uplink burst, expressed in analog-to-digital converter output code value. This represents median operations; This represents the set of sampling point numbers at the previous uplink sudden shutdown location.

[0044] It should be noted that the residual at the end of the previous uplink burst is only used to determine the initial residual value and residual polarity, and is not directly used for exponential fitting with the guard interval sampling signal, so as to avoid mistaking the effective communication component that has not yet ended the previous uplink burst as a free decay response.

[0045] 1.2 Reconstruct the current upward burst precursor signal.

[0046] Based on the known preamble sequence, coarse timing results, and preliminary amplitude estimation results of the current uplink burst, the expected communication signal of the current uplink burst preamble is reconstructed as follows:

[0047] ;

[0048] in, Indicates the current upward emergency leading segment. The reconstructed signal of each sampling point is expressed in units of analog-to-digital converter output code value. This indicates the sampling point number of the current uplink burst leader segment; This represents the initial amplitude estimate of the current uplink burst, expressed in analog-to-digital converter output code values. Indicates the sequence number of the preamble symbol; The number of preamble symbols involved in the reconstruction is a dimensionless integer. Indicates the first There are 1 known preamble code elements, which are dimensionless values; This represents the normalized nominal impulse response obtained from receiver calibration, which is a dimensionless function. This indicates the current coarse timing offset of the uplink burst, in seconds.

[0049] The reconstructed signal is subtracted from the original sampled signal of the current uplink burst preamble to obtain the preamble residual, which is expressed as:

[0050] ;

[0051] in, Indicates the current upward emergency leading segment. The residual of each sampling point, in units of analog-to-digital converter output code value; Indicates the current upward emergency leading segment. The original sampled values ​​of each sampling point are expressed in units of analog-to-digital converter output code values.

[0052] 1.3 Forming an observation sequence.

[0053] The residuals of the guard interval sampling signal and the current uplink burst leader are arranged in actual time order to form an observation sequence for recovery response estimation, and the elapsed time of each observation sampling point relative to the end time of the previous uplink burst is recorded.

[0054] Furthermore, local median deviation detection is performed on the observed sequence.

[0055] When the difference between a sampling point and its local median exceeds the anomaly detection threshold, the sampling point exceeding the anomaly detection threshold is marked as a low-confidence sampling point, but the current sampling point is not directly deleted.

[0056] The anomaly detection threshold is obtained by calibrating the absolute deviation of the local median of the sampled signal in the protection interval without anomalies, and is usually taken as 3 to 6 times the absolute deviation of the local median.

[0057] 1.4 Calculate the amplitude normalization benchmark.

[0058] The amplitude difference is calculated based on the average sampling amplitude before the end of the previous uplink burst and the baseline sampling amplitude at the rear of the protection interval, and is expressed as follows:

[0059] ;

[0060] in, This represents the difference between the amplitude of the previous uplink burst termination and the amplitude of the guard interval baseline, expressed in analog-to-digital converter output code value. This represents the average sampling amplitude within the sampling window before the end of the previous uplink burst, expressed in analog-to-digital converter output code values. This represents the average sampling amplitude of the baseline within the sampling window after the guard interval, expressed in analog-to-digital converter output code values.

[0061] Furthermore, the complete original sampled signal of the current uplink burst is written into the original sampling buffer. The digital buffer stores at least all sampled values ​​of the current uplink burst from the start position of the preamble to the end position of the data segment.

[0062] It should be noted that the digital cache is used for the full compensation process in S5, and for the uncompensated rollback process in S6 when the compensation result fails to pass verification.

[0063] Finally, S1 outputs the observation sequence, the elapsed time of each observation sampling point, the initial residual value, the amplitude difference, the protection interval time, the current uplink burst original sampling buffer, and the power difference between the previous and subsequent uplink bursts.

[0064] S2. Based on the receiver calibration results and the operating status of adjacent uplink bursts, select candidate time constants, establish an exponential basis function matrix, and perform detrending and amplitude normalization on the observation sequence.

[0065] Before the receiver is put into operation, test burst optical signals with different power, duration and gain are input to the optical receiver, and the recovery curve after the test burst ends is collected.

[0066] Furthermore, based on the time required for the sampled amplitude in the recovery curve to decrease to a preset proportion of the initial residual value, the candidate time constant range for the receiver's recovery response is determined.

[0067] The candidate time constants are set according to logarithmic intervals, expressed as follows:

[0068] ;

[0069] in, Indicates the first One candidate time constant, in seconds; This indicates the lower limit of the candidate time constant, in seconds. This indicates the upper limit of the candidate time constant, in seconds; Indicates the candidate time constant index; This represents the total number of candidate time constants, which is a dimensionless integer.

[0070] The lower and upper limits of the candidate time constant are determined based on the receiver step response and the recovery curve after the test burst shutdown. The lower limit of the candidate time constant is usually 0.5 to 5 symbol periods, and the upper limit of the candidate time constant is usually 20 to 500 symbol periods.

[0071] Furthermore, based on the difference in uplink burst power between the previous and subsequent bursts, the duration of the previous uplink burst, the length of the guard interval, and the current receiver gain status, a candidate time constant for the current processing cycle is selected from all candidate time constants.

[0072] When the recovery trajectory within the protection interval mainly exhibits a rapid decline, the fast and medium-speed candidate time constants are retained; when there is still continuous baseline change at the end of the protection interval, the slow-speed candidate time constant is also retained.

[0073] Furthermore, the criteria for judging rapid decline and continuous baseline changes are as follows: the protection interval is divided into the first, middle, and last segments according to time sequence, and the amplitude difference obtained by S1 is used as the criterion. As a normalization benchmark; when the ratio of the median absolute value of the residuals in the first segment of the protection interval to the median absolute value of the residuals in the second segment is not less than 2, and the time taken for the residual envelope to decrease from 90% of the initial value to 10% does not exceed 20% of the protection interval time, then the recovery trajectory within the protection interval is judged to be rapidly decreasing; when the product of the absolute value of the baseline slope obtained by robust linear regression in the second segment of the protection interval and the duration of the second segment accounts for If the ratio is not less than 0.05, and at least 80% of the adjacent sampling differences in the latter segment are in the same direction as the baseline slope, then it is determined that there is a continuous baseline change at the end of the protection interval. Among them, the judgment value of the residual ratio between the front and rear segments is usually taken as 2 to 5, and the normalized baseline change ratio is usually taken as 0.03 to 0.10. The specific values ​​are determined based on the calibration data of no obvious recovery samples, fast recovery samples, and slow baseline recovery samples.

[0074] Based on the elapsed time of each observation sampling point relative to the end time of the previous uplink burst, an exponential basis function matrix is ​​established, represented as:

[0075] ;

[0076] in, The exponential basis function matrix represents the first... Line 1 The elements of the column are dimensionless numerical values; Indicates the first The elapsed time of each observation sampling point relative to the end of the previous uplink burst, in seconds; Indicates the sequence number of the observation sampling point; This represents the natural exponential function.

[0077] Furthermore, the correlation between each candidate exponential basis function and the constant baseline and the linear baseline is calculated.

[0078] When a certain long-time constant exponential basis function is highly correlated with a constant baseline or a linear baseline, the current exponential basis function is removed from the exponential basis function matrix to prevent the slow exponential response from replacing the baseline drift.

[0079] 2.1 Detrending and amplitude normalization are performed on the observed sequence.

[0080] Robust linear regression is used to estimate the baseline constant term and baseline rate of change within the observation window, and the observation series is detrended and amplitude normalized, as shown below:

[0081] ;

[0082] in, Indicates the first Each normalized observation is a dimensionless value. Indicates the first One raw observation value, in units of analog-to-digital converter output code value; This represents the estimated baseline constant term, expressed in analog-to-digital converter output code values. This represents the estimated baseline change rate, expressed in analog-to-digital converter output code values ​​per second.

[0083] It should be noted that the amplitude stabilization value can be taken as 0.5 to 1 times the root mean square value of the analog-to-digital converter sampling noise when there is no signal input, in order to avoid abnormally large normalization results when the difference between the amplitudes is small.

[0084] Furthermore, the initial residual values ​​are normalized according to the same amplitude benchmark, and are expressed as:

[0085] ;

[0086] in, This represents the initial residual value after normalization, which is a dimensionless value.

[0087] Finally, S2 outputs the exponential basis function matrix, the candidate time constant set, and the normalized observation sequence.

[0088] S3. Based on the exponential basis function matrix, perform sparse estimation with amplitude boundaries and initial continuity constraints on the normalized observation sequence to obtain candidate memory components.

[0089] Based on the difference in uplink burst power, the length of the protection interval, and the receiver gain status, query historical processing records and read historical coefficients saved under the same operating conditions.

[0090] Furthermore, the historical coefficients read are used as the initial values ​​for the current sparse estimation iteration; if no historical coefficients exist under the same operating conditions, the initial values ​​of all candidate coefficients are set to zero.

[0091] A sparse estimation model with observation weights, magnitude boundaries, and initial continuity constraints is adopted, and is expressed as follows:

[0092] ;

[0093] The constraints satisfied by the sparse estimation model are as follows: and .

[0094] in, This represents the candidate coefficient vector obtained from the sparse estimation; it is a dimensionless vector. represents the candidate coefficient vector to be determined, which is a dimensionless vector; This represents the observation vector composed of all normalized observations, and is a dimensionless vector. represents the exponential basis function matrix, which is a dimensionless matrix; This represents the observation weights diagonal matrix, which is a dimensionless matrix. represents the sparse regularization coefficient, which is a dimensionless value; Represents the L2 norm; Denotes the norm 1; Indicates the first There are 1 candidate coefficients, which are dimensionless values; Indicates the first The upper limit of the amplitude of each candidate coefficient is a dimensionless value; superscript Indicates vector transpose; This represents the initial residual value after normalization; This represents the initial continuity threshold, which is a dimensionless value.

[0095] The initial continuity threshold is obtained by jointly calibrating the reconstruction error at the end of the previous uplink burst and the normalized sampling noise, and is usually taken as 0.03 to 0.15.

[0096] The upper limit of the candidate coefficient amplitude is determined based on the maximum ratio of the recovery response amplitude difference relative to the previous uplink burst termination amplitude at different time scales, and is usually taken as 0.10 to 1.50.

[0097] The sparse regularization coefficient is determined based on the cross-validation results between the number of fitting residuals and the number of error components in the independently calibrated data, and is usually taken as 0.001 to 0.08 after normalization.

[0098] Furthermore, the observation weight of the normal sampling points in the protection interval is set to 1, the observation weight of the residual sampling points in the current uplink burst preamble is set to 0.2 to 0.6 based on the preamble signal reconstruction error, and the observation weight of the abnormal sampling points is set to 0.05 to 0.30.

[0099] A fast iterative shrinking threshold algorithm is used to solve the sparse estimation model, and the solution is terminated when the change in candidate coefficients in two consecutive iterations is less than the stopping condition or the maximum number of iterations is reached.

[0100] Furthermore, candidate coefficients with absolute values ​​lower than the noise amplitude threshold are set to zero, and the remaining non-zero candidate coefficients are combined with the corresponding candidate time constants to obtain candidate memory components.

[0101] The noise amplitude threshold is obtained by calibrating the standard deviation of the normalized sampling noise under the condition of no uplink signal, and is usually taken as 2.5 to 4.5 times the standard deviation of noise.

[0102] Calculate the normalized residual energy when using all candidate memory components and the normalized baseline residual energy when not using candidate memory components, and pass the candidate coefficients, corresponding time constants, and the two residual energies to S4.

[0103] S4. Merge candidate memory components with the same sign and the closest correlation based on the correlation of the exponential basis functions, and select effective memory components based on residual contribution, cross-window stability and fitting residual ratio.

[0104] 4.1 Merge candidate memory components with the same and closest numbers.

[0105] The normalized correlation coefficient between any two candidate exponential basis functions is calculated as follows:

[0106] ;

[0107] in, Indicates the first The exponential basis function and the first The normalized correlation coefficient between the exponential basis functions is a dimensionless value; and Indicates two different candidate time constant indices; and These represent the nth element in the exponential basis function matrix. Column and number List.

[0108] The component similarity threshold is determined based on the distribution of normalized correlation coefficients of different exponential basis functions within the actual observation time window, and is usually taken as 0.95 to 0.995.

[0109] When the normalized correlation coefficients of two candidate exponential basis functions reach the component similarity threshold and the corresponding candidate coefficients have the same sign, the corresponding candidate memory components are grouped into the same group.

[0110] For the same group, the time constant corresponding to the candidate coefficient with the largest absolute value is retained as the equivalent time constant.

[0111] After summing the amplitudes of all candidate coefficients within the same group, the analog-to-digital converter output code value is recovered based on the amplitude normalization reference obtained in S1, thus obtaining the equivalent amplitude of the candidate memory components in that group; the equivalent time constant of that group is taken as... .

[0112] Furthermore, when two highly correlated candidate coefficients have opposite signs, they are not merged; instead, subsequent residual contribution judgments are performed separately to prevent the erroneous merging of positive overshoot and negative swirl.

[0113] 4.2 Calculate the residual contribution.

[0114] Remove each candidate memory component after grouping in turn, while keeping the other candidate memory components unchanged, and recalculate the normalized residual energy.

[0115] No. The residual contribution of each candidate memory component is expressed as:

[0116] ;

[0117] in, Indicates the first The residual contribution of each candidate memory component is a dimensionless value. This indicates the sequence number of the candidate memory component after the grouping process is completed; Indicates the removal of the first The normalized residual energy after each candidate memory component is a dimensionless value. This represents the normalized residual energy when using all candidate memory components; it is a dimensionless value. This represents the normalized baseline residual energy when candidate memory components are not used; it is a dimensionless value. It represents the residual stabilizing quantity to prevent the denominator from being zero, and is a dimensionless positive number; and These represent taking the minimum and maximum values, respectively.

[0118] The contribution threshold is determined based on the residual contribution distribution of the true recovery component and the noise fitting component, and is typically taken as 0.05 to 0.20.

[0119] 4.3 Calculate cross-window stability.

[0120] The observation sequence is divided into multiple overlapping sub-windows, and local sparse estimation is performed within each sub-window using the candidate exponential basis functions determined by S3.

[0121] No. The cross-window stability of each candidate memory component is expressed as:

[0122] ;

[0123] in, Indicates the first The cross-window stability of each candidate memory component is a dimensionless value. Indicates the first The number of times each candidate memory component is retained in the sub-window estimation is a dimensionless integer; This represents the total number of child windows and is a dimensionless integer.

[0124] The stability threshold is determined based on the retention ratio of the true recovery component in the overlapping sub-window, and is typically between 0.60 and 0.85.

[0125] Furthermore, the ratio between the normalized residual energy of all candidate memory components and the normalized baseline residual energy is calculated.

[0126] The fitting residual ratio threshold is determined based on the statistical results of the normalized residual ratio of valid trail samples and samples without valid trails, and is usually taken as 0.10 to 0.35.

[0127] When the residual contribution of a candidate memory component reaches the contribution threshold, the cross-window stability reaches the stability threshold, the fitting residual ratio does not exceed the fitting residual ratio threshold, and the equivalent amplitude is not lower than the noise amplitude threshold, the current candidate memory component is determined as a valid memory component.

[0128] Finally, S4 outputs the equivalent amplitude and equivalent time constant of the effective memory component, as well as the position of the effective memory component in the original exponential basis function matrix.

[0129] S5. Based on the effective memory components, predict the residual amount in the current uplink burst, use multiple candidate compensation ratios to process the current uplink burst leader segment, and determine the target compensation ratio.

[0130] Based on the correlation results between the current uplink burst preamble and the original sampled signal, an interpolation correlation search is performed at the integer sampling position to obtain the fractional sampling level timing correction.

[0131] No. The effective memory components are in the current upward burst. The predicted residual amount at each sampling point is expressed as:

[0132] ;

[0133] in, Indicates the first The effective memory components are in the current upward burst. The predicted residual value for each sampling point, in units of analog-to-digital converter output code value; Indicates the first The equivalent amplitude of each effective memory component, in units of analog-to-digital converter output code value; This represents the time from the end of the previous uplink burst to the nominal start of the current uplink burst, in seconds. Indicates the sampling point number within the current uplink burst; Indicates the sampling period, in seconds; This indicates the fractional sampling level timing correction, in seconds. Indicates the first The equivalent time constant of each effective memory component, in seconds.

[0134] The predicted residuals of all valid memory components at the same sampling point are added together to obtain the total predicted residuals of all sampling points of the current uplink burst.

[0135] Furthermore, multiple candidate compensation ratios are set between 0 and 1, where 0 corresponds to no compensation and 1 corresponds to full deduction of the total predicted residual amount.

[0136] The candidate compensation ratio is obtained based on the receiver dynamic range and residual prediction error calibration, and the value ranges from 0 to 1. The interval between adjacent candidate compensation ratios is usually 0.1 to 0.25.

[0137] Adopting the first Each candidate compensation ratio is used to process the current uplink burst leader segment, and is expressed as follows:

[0138] ;

[0139] in, Indicates the use of the first After the first candidate compensation ratio The signal at each sampling point is represented by the analog-to-digital converter output code value. Indicates the current upward sudden event. The original sampled values ​​of each sampling point are expressed in units of analog-to-digital converter output code values. Indicates the first Each candidate compensation ratio is a dimensionless value. Indicates the candidate compensation ratio sequence number; This indicates that all effective memory components are in the first... The total predicted residual for each sampling point, expressed in analog-to-digital converter output code value.

[0140] For each candidate compensation result, the normalized preamble correlation quality, normalized high-low level spacing, normalized DC residual, and normalized initial equalization error are calculated respectively.

[0141] Furthermore, candidate compensation ratios that do not meet the improvement requirements related to the preamble are first excluded.

[0142] The preamble correlation improvement threshold is determined based on the relationship between the amount of preamble correlation quality improvement after compensation and the synchronization success rate, and is usually taken as 0.02 to 0.10.

[0143] Furthermore, among the remaining candidate compensation ratios, those whose high-low level spacing degradation exceeds the allowable range are excluded.

[0144] The degradation threshold for the high and low level spacing is determined based on the fluctuation range of the high and low level spacing under normal receiving conditions, and is usually taken as 0.01 to 0.08.

[0145] Furthermore, candidate compensation ratios with normalized DC residuals exceeding the allowable range are excluded from the remaining candidate compensation ratios.

[0146] The DC residual threshold is determined based on the normalized leader DC residual distribution in the successfully decoded samples, and is typically taken as 0.05 to 0.20.

[0147] Among the candidate compensation ratios that satisfy the constraints of preamble correlation quality, high and low level spacing, and DC residual, the candidate compensation ratio with the smallest normalized initial equalization error is selected as the target compensation ratio.

[0148] If there is no candidate compensation ratio that meets the above conditions, the ratio for which no compensation is performed will be determined as the target compensation ratio.

[0149] Read the complete original sampled signal of the current uplink burst from the original sampling buffer formed by S1, subtract the total prediction residual according to the target compensation ratio, and perform gain adjustment, clock recovery, adaptive equalization and decoding in sequence.

[0150] S6. Process the current uplink burst according to the target compensation ratio. If the verification is successful, output the communication data and update the historical coefficients. If the verification fails, call the original sampling buffer to perform uncompensated processing.

[0151] Determine whether the current uplink burst, after being processed using the target compensation ratio, meets the following compensation processing conditions:

[0152] First, the adaptive equalizer converges within a specified number of iterations.

[0153] Second, the equilibrium residual is within the residual range corresponding to the normal operating sample of the receiver.

[0154] Third, forward error correction decoding or frame verification passes.

[0155] When the target compensation processing meets the compensation processing conditions, the decoding result after compensation processing will be used as the final communication data of the current uplink burst.

[0156] Furthermore, based on the position of the effective memory component in the original exponential basis function matrix during the current processing cycle, a fixed-dimensional set of current effective coefficients is formed; candidate coefficients that fail the validity judgment are set to zero.

[0157] The historical coefficients under the same operating conditions are updated based on the current effective coefficients, expressed as follows:

[0158] ;

[0159] in, Indicates the first In the processing cycle, the first The historical coefficients corresponding to each exponential basis function are dimensionless values. Indicates the first In the processing cycle, the first The historical coefficients corresponding to each exponential basis function are dimensionless values. Indicates the first The first one that passed the validity judgment in the processing cycle There are 1 candidate coefficients, which are dimensionless values. If a coefficient fails the validity judgment, it is set to 0. Indicates the processing cycle number; Indicates the candidate time constant index; This represents the forgetting coefficient, which is a dimensionless value.

[0160] The forgetting factor is determined based on the rate of change of receiver temperature and gain state, and is typically taken as 0.80 to 0.98.

[0161] It should be noted that the updated historical coefficients are only used as the initial values ​​for the sparse estimation under the same operating conditions in the next iteration, and do not directly replace the real-time estimation results in the next processing cycle.

[0162] When the target compensation process fails any of the equalization convergence, equalization residual, or decoding verification, the original sampled signal of the current uplink burst is read from the digital buffer, and regular gain adjustment, clock recovery, adaptive equalization, and decoding are performed without deducting the prediction residual.

[0163] When the uncompensated processing passes the decoding verification, the decoding result of the uncompensated processing is used as the final communication data of the current uplink burst, and the use of the current candidate coefficients to update the historical coefficients is prohibited.

[0164] When both the compensated and uncompensated processing fail the decoding verification, the current uplink burst reception failure flag is output, and the historical coefficients are not updated.

[0165] Furthermore, regardless of whether compensation processing or no compensation processing is ultimately adopted, the decision symbols, amplitude estimation results, equivalent received impulse response, timing offset, and current uplink burst end position obtained during the final effective processing are all preserved.

[0166] This embodiment also provides a computer device applicable to optical signal processing methods in passive optical networks, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the optical signal processing method in a passive optical network as proposed in the above embodiment.

[0167] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0168] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the optical signal processing method in the passive optical network proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0169] In summary, this invention achieves accurate characterization of composite recovery wakes by constructing a multi-timescale exponential basis function matrix; achieves effective separation of recovery response from noise and baseline drift by constrained sparse estimation and effective component screening; achieves adaptive determination of compensation degree by comparing candidate compensation ratios and constraining compensation quality; and achieves reliable reception of uplink burst optical signals by joint verification of compensation results and back-off of the original sampling buffer.

[0170] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for optical signal processing in a passive optical network, characterized in that, include: Collect digital sampling signals from the end of the previous uplink burst, the guard interval, and the leading segment of the current uplink burst; remove reconfigurable communication signals to obtain the observation sequence and the original sampling buffer of the current uplink burst. Candidate time constants are selected based on receiver calibration results and the operating status of adjacent uplink bursts, an exponential basis function matrix is ​​established, and the observation sequence is detrended and amplitude normalized. Based on the exponential basis function matrix, sparse estimation with amplitude boundary and initial continuity constraints is performed on the normalized observation sequence to obtain candidate memory components; Candidate memory components with the same sign and the closest correlation are merged based on the correlation of the exponential basis functions, and effective memory components are selected based on residual contribution, cross-window stability and fitting residual ratio. Based on the effective memory components, the residual amount in the current uplink burst is predicted. Multiple candidate compensation ratios are used to process the current uplink burst preamble. The target compensation ratio is determined based on the preamble correlation quality, DC residual and initial equalization error. The current uplink burst is processed according to the target compensation ratio, and gain adjustment, clock recovery, adaptive equalization and decoding are performed in sequence. When the verification is successful, communication data is output and historical coefficients are updated. When the verification fails, the original sampling buffer is called to perform uncompensated processing.

2. The optical signal processing method in a passive optical network as described in claim 1, characterized in that, The specific steps for obtaining the observation sequence and the original sampling buffer of the current uplink burst are as follows: Based on the uplink authorization time, the edge of the digital sampling amplitude change, and the preamble, determine the end position of the previous uplink burst, the protection interval, and the coarse start position of the current uplink burst; Based on the decision symbols, amplitude estimates, equivalent received impulse responses and timing offsets stored in the previous uplink burst, the expected communication signal at the end of the previous uplink burst is reconstructed, and the expected communication signal is subtracted from the original sampled signal to obtain the end residual and the initial residual value. Based on the known preamble sequence, coarse timing, and preliminary amplitude estimation of the current uplink burst, the expected communication signal of the current uplink burst preamble is reconstructed, and the expected communication signal is subtracted from the original sampled signal of the preamble to obtain the preamble residual. The guard interval sampling signal and the leader segment residual are arranged in the actual time sequence to form an observation sequence, and the complete original sampling signal of the current uplink burst is written into the original sampling buffer.

3. The optical signal processing method in a passive optical network as described in claim 1, characterized in that, The specific steps for establishing the exponential basis function matrix and performing detrending and magnitude normalization on the observed sequence are as follows: The range of candidate time constants is determined based on the receiver step response and the recovery curve after the test burst ends, and the candidate time constants are set at logarithmic intervals within the range of candidate time constants; Based on the difference in uplink burst power between the previous and next uplink bursts, the duration of the previous uplink burst, the guard interval length, and the current receiver gain status, the candidate time constant to be used in the current processing cycle is selected from the candidate time constants. Based on the elapsed time of each observation sampling point in the observation sequence relative to the end time of the previous uplink burst and the selected candidate time constant, an exponential basis function matrix is ​​established, and exponential basis functions that reach a preset degree with the constant baseline or linear baseline are deleted. Robust linear regression is used to estimate the baseline constant term and baseline rate of change of the observed sequence. Based on the amplitude difference between the amplitude of the previous uplink burst termination and the baseline amplitude of the protection interval, the observed sequence is detrended and amplitude normalized.

4. The optical signal processing method in a passive optical network as described in claim 1, characterized in that, The process of performing sparse estimation with amplitude boundaries and initial continuity constraints on the normalized observation sequence to obtain candidate memory components involves the following steps: Based on the difference in uplink burst power, the guard interval length, and the receiver gain status, the historical coefficients stored under the same operating conditions are read, and the historical coefficients are used as the initial values ​​for the sparse estimation iteration. Set up an observation weight matrix, where the normal sampling points of the protection interval, the residual sampling points of the current uplink burst leader segment, and the low-confidence sampling points each correspond to different observation weights. The optimization targets are the weighted fitting error between the candidate coefficient vector and the normalized observation sequence and the first norm of the candidate coefficient vector. Amplitude boundaries and initial continuity constraints corresponding to the normalized initial residual values ​​are set for each candidate coefficient. A fast iterative shrinkage threshold algorithm is used to solve for candidate coefficients. Candidate coefficients with absolute values ​​lower than the noise amplitude threshold are set to zero, and the remaining candidate coefficients are combined with the corresponding candidate time constants to form candidate memory components.

5. The optical signal processing method in a passive optical network as described in claim 1, characterized in that, The specific steps for merging the most similar candidate memory components with the same sign based on the correlation of the exponential basis function are as follows: Calculate the normalized correlation coefficient between the exponential basis functions of any two candidate memory components; When the normalized correlation coefficient between two exponential basis functions reaches the component similarity threshold and the corresponding candidate coefficients have the same sign, the corresponding candidate memory components are grouped into the same group. For the same group, the candidate time constant corresponding to the candidate coefficient with the largest absolute value is determined as the equivalent time constant, and the amplitudes of each candidate coefficient in the group are added together and restored to the equivalent amplitude according to the amplitude normalization benchmark. When the normalized correlation coefficient between two exponential basis functions reaches the component similarity threshold, and the corresponding candidate coefficients have opposite signs, the corresponding candidate memory components are retained for subsequent screening.

6. The optical signal processing method in a passive optical network as described in claim 1 or 5, characterized in that, The specific steps for selecting effective memory components based on residual contribution, cross-window stability, and fitted residual ratio are as follows: Remove each candidate memory component sequentially while keeping the remaining candidate memory components unchanged. Calculate the normalized residual energy before and after removal, and determine the residual contribution of each candidate memory component based on the change in normalized residual energy. The observation sequence is divided into at least two overlapping sub-windows. Local sparsity estimation is performed within each sub-window based on candidate exponential basis functions. Cross-window stability is determined based on the number of times the candidate memory component is retained in at least two sub-windows. Calculate the ratio of the fitted residual energy between the normalized residual energy using all candidate memory components and the normalized baseline residual energy without using candidate memory components; Candidate memory components whose residual contribution reaches the contribution threshold, whose cross-window stability reaches the stability threshold, whose fitted residual ratio does not exceed the fitted residual ratio threshold, and whose equivalent amplitude is not lower than the noise amplitude threshold are identified as effective memory components.

7. The optical signal processing method in a passive optical network as described in claim 1, characterized in that, The specific steps for determining the target compensation ratio based on the preamble correlation quality, DC residual, and initial equalization error are as follows: Based on the current uplink burst preamble and the original sampled signal, perform interpolation correlation search at integer sampling positions to obtain fractional sampling level timing correction. Based on the equivalent amplitude, equivalent time constant, protection interval time and fractional sampling level timing correction of each effective memory component, the residual amount of each effective memory component at each sampling point of the current uplink burst is predicted, and the residual amounts at the same sampling points are added together to obtain the total predicted residual amount. Set at least two candidate compensation ratios between zero and one, and subtract the total predicted residual from the original sampled signal of the current uplink burst leader according to each candidate compensation ratio; The normalized preamble correlation quality, normalized high-low level spacing, normalized DC residual, and normalized initial equalization error of each candidate compensation result are calculated respectively. Candidate compensation ratios that do not meet the preamble correlation improvement requirements, have high-low level spacing degradation exceeding the allowable range, and have normalized DC residual exceeding the allowable range are eliminated in turn. Among the remaining candidate compensation ratios, the candidate compensation ratio with the smallest normalized initial equilibrium error is determined as the target compensation ratio. If there is no candidate compensation ratio that meets the conditions, the ratio for which compensation is not performed is determined as the target compensation ratio.

8. The optical signal processing method in a passive optical network as described in claim 1, characterized in that, The process of handling the current uplink burst according to the target compensation ratio involves sequentially performing gain adjustment, clock recovery, adaptive equalization, and decoding. The specific steps are as follows: Read the complete original sampled signal of the current uplink burst from the original sampling buffer, subtract the total predicted residual amount according to the target compensation ratio, and perform gain adjustment, clock recovery, adaptive equalization and decoding in sequence; The compensation processing results are verified based on whether the adaptive equalizer converges within the specified number of iterations, whether the equalization residual is within the residual range corresponding to the normal working sample of the receiver, and whether the forward error correction decoding or frame check passes. When the compensation processing result passes the verification, the communication data after compensation processing is output. The current effective coefficient is formed according to the position of the effective memory component in the exponential basis function matrix, and the historical coefficient under the same operating conditions is updated using the forgetting coefficient. When the compensation processing result fails verification, the original sampling buffer is called to perform uncompensated processing without deducting the total prediction residual. When the uncompensated processing passes the decoding verification, the communication data after uncompensated processing is output and the historical coefficients are not updated. When both the compensation processing and the uncompensated processing fail the decoding verification, the current uplink burst reception failure flag is output.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the optical signal processing method in the passive optical network according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the optical signal processing method in the passive optical network according to any one of claims 1 to 8.

Images