A forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS coding

Abstract

The application discloses a 1.6T PAM4 signal forward error correction method based on RS coding optimal check bit number, relates to the technical field of communication forward error correction, and comprises the following steps: collecting link observation data of 1.6T PAM4 signals, establishing an RS code word level error observation window according to RS code word boundary information, obtaining the number of error symbols, burst error coverage information and error code rate of each RS code word before error correction, calculating the maximum number of error symbols, error symbol number discrete degree and high error code word proportion, generating a code word internal error aggregation index, and determining an aggregation correction margin in combination with the burst error coverage information; screening candidate RS coding check bit numbers according to a target error code rate after error correction, the maximum number of error symbols and the aggregation correction margin, selecting the one with the lowest coding overhead as the RS coding optimal check bit number, and performing RS forward error correction coding and decoding. The application realizes error code sensing and RS coding optimal check bit number selection, and takes into account error correction reliability and transmission overhead.

Description

Technical Field

[0001] This invention relates to the field of communication forward error correction technology, and in particular to a forward error correction method for 1.6TPAM4 signals based on the optimal parity bit length of RS coding. Background Technology

[0002] As data centers and high-speed Ethernet evolve towards 1.6Tb / s, four-level pulse amplitude modulation (PAM4) has become the mainstream signal modulation method due to its high bandwidth efficiency. However, PAM4 signals are more sensitive to noise and channel nonlinearity due to their multi-level decision mechanism, leading to an increased link bit error rate. To ensure transmission reliability, the industry generally adopts Reed-Solomon (RS) codes as a forward error correction (FEC) scheme. Conventional RS-FEC methods are based on a preset, fixed parity bit configuration, such as based on bit error statistics under worst-case channel conditions or the error correction capabilities specified by standard protocols, providing uniform and static error correction protection for the entire data stream.

[0003] However, the static parity bit configuration is difficult to adapt to the dynamic time-varying characteristics of the channel state in a 1.6T PAM4 link. On the one hand, a fixed parity bit introduces unnecessary coding overhead and transmission delay when the channel quality is good, while its preset error correction capability may be insufficient when there is sudden strong interference or when error symbols are highly concentrated in a few codewords, leading to decoding failure. On the other hand, traditional methods rely only on the average bit error rate before error correction to evaluate the link state, which fails to fully characterize the clustering distribution characteristics of error symbols within the codeword and the degree of local contamination of sudden errors, resulting in a mismatch risk between the allocation of redundant resources and the actual error distribution. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS coding to solve the problems of poor dynamic adaptability of coding overhead to channel state and insufficient response to the risk of clustering of erroneous symbols within codewords and local contamination of sudden bit errors in existing technologies.

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

[0007] This invention provides a forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS encoding, comprising: acquiring 1.6T... Based on the link observation data of the PAM4 signal, an RS codeword-level error observation window is established according to the RS codeword boundary information to obtain the number of erroneous symbols, burst error coverage information, and pre-correction bit error rate for each RS codeword. Based on the number of erroneous symbols in each RS codeword, the maximum number of erroneous symbols, the dispersion of erroneous symbols, and the proportion of high-error codewords are calculated to generate an intra-codeword error clustering index. Based on the intra-codeword error clustering index and burst error coverage information, the clustering correction margin is determined. Based on the target post-correction bit error rate, the pre-correction bit error rate, the candidate RS code parity bit set, the maximum number of erroneous symbols, and the clustering correction margin, candidate RS code parity bits are selected where the estimated post-correction bit error rate is not higher than the target post-correction bit error rate, and the number of correctable symbols is not less than the sum of the maximum number of erroneous symbols and the clustering correction margin. Among the selected candidate RS code parity bits, the candidate RS code parity bit with the lowest coding overhead is selected as the optimal RS code parity bit, and the optimal RS code parity bit is used to adjust the 1.6T... The PAM4 data stream undergoes RS forward error correction coding and decoding, outputting a 1.6TPAM4 data stream after error correction.

[0008] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS coding described in this invention, the step of obtaining the number of erroneous symbols, burst error coverage information, and bit error rate before error correction for each RS codeword includes: performing PAM4 level decision on the received signals of each high-speed PAM4 physical channel; converting the obtained PAM4 decision symbol stream into a bit data stream; determining the starting position of the RS codeword based on the codeword synchronization marker in the bit data stream; dividing the RS codewords according to their length order based on the determined starting position to obtain RS codeword boundary information; performing a syndrome calculation on each RS codeword; and determining the position of the erroneous symbol based on the error localization polynomial when the syndrome calculation result is not zero. The number of erroneous symbols in each RS codeword is counted based on the error markers of the RS symbols. RS codewords continuously received within one observation period on the same high-speed PAM4 physical channel are divided into RS codeword-level error observation windows. Adjacent erroneous symbols with an error interval less than the burst interval threshold are grouped into the same burst error event. Based on the mapping relationship between burst error events and RS codeword boundaries, the number of RS codewords covered by burst error events and the number of erroneous symbols within each covered RS codeword are counted to obtain burst error coverage information. The bit error rate before correction is calculated based on the ratio of the number of erroneous bits detected before RS error correction to the total number of received bits within the RS codeword-level error observation window.

[0009] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS coding described in this invention, the specific steps for calculating the maximum number of erroneous symbols, the dispersion of the number of erroneous symbols, and the proportion of high-erroneous codewords based on the number of erroneous symbols in each RS codeword are as follows: traversing the codeword-level error sequence corresponding to the current RS codeword-level error observation window; determining the maximum value in the codeword-level error sequence as the maximum number of erroneous symbols; calculating the average number of erroneous symbols for each RS codeword within the current RS codeword-level error observation window, and determining the dispersion of the number of erroneous symbols based on the degree of deviation of the number of erroneous symbols for each RS codeword from the average number of erroneous symbols; determining RS codewords with a number of erroneous symbols greater than or equal to the error concern threshold as high-erroneous codewords; and determining the proportion of high-erroneous codewords based on the proportion of the number of high-erroneous codewords to the total number of RS codewords within the current RS codeword-level error observation window.

[0010] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding described in this invention, the generated codeword error clustering index includes: calculating the difference between the maximum number of erroneous symbols and the average number of erroneous symbols, and normalizing the difference using the average number of erroneous symbols to obtain the maximum error deviation term; normalizing the dispersion of the number of erroneous symbols using the average number of erroneous symbols to obtain the dispersion term; adjusting the maximum error deviation term, the dispersion term, and the high error codeword proportion term according to the corresponding adjustment coefficients, and superimposing the adjusted maximum error deviation term, dispersion term, and high error codeword proportion term to obtain the codeword error clustering index.

[0011] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit depth of RS coding described in this invention, the determination of the aggregation correction margin includes: for each burst error event within the current RS codeword-level error observation window, counting the number of RS codewords covered by the burst error event and the number of error symbols within each covered RS codeword; determining the burst error concentration degree corresponding to each burst error event based on the maximum number of error symbols, the total number of error symbols, and the number of covered RS codewords for each burst error event, and determining the maximum value among the burst error concentration degrees corresponding to all burst error events as the burst error concentration degree of the current RS codeword-level error observation window; adjusting the error aggregation index within the codeword and the burst error concentration degree according to the corresponding adjustment coefficients, and superimposing the adjusted error aggregation index within the codeword and the burst error concentration degree to obtain the aggregation risk value.

[0012] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS coding described in this invention, the determination of the aggregation correction margin further includes: comparing the aggregation risk value with a first aggregation risk threshold, a second aggregation risk threshold, and a third aggregation risk threshold, respectively; when the aggregation risk value is less than the first aggregation risk threshold, the aggregation correction margin is determined to be zero RS symbols; when the aggregation risk value is not less than the first aggregation risk threshold and is less than the second aggregation risk threshold, the aggregation correction margin is determined to be one RS symbol; when the aggregation risk value is not less than the second aggregation risk threshold and is less than the third aggregation risk threshold, the aggregation correction margin is determined to be two RS symbols; when the aggregation risk value is not less than the third aggregation risk threshold, the aggregation correction margin is determined to be three RS symbols.

[0013] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS coding described in this invention, the specific steps for selecting candidate RS coding parity bit lengths that have an estimated post-correction bit error rate not higher than the target post-correction bit error rate and a number of correctable symbols not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin are as follows: Based on the number of bits contained in each RS symbol, each candidate RS coding parity bit length is converted into a number of redundant parity symbols, and the number of correctable symbols is determined based on the number of redundant parity symbols; the sum of the maximum number of erroneous symbols and the aggregation correction margin is determined as the minimum correctable symbol requirement; the candidate RS coding parity bit lengths with a number of correctable symbols not less than the minimum correctable symbol requirement are determined as preliminary candidate RS coding parity bit lengths, and candidate RS coding parity bit lengths with a number of correctable symbols less than the minimum correctable symbol requirement are eliminated.

[0014] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS coding described in this invention, the specific steps of selecting candidate RS coding parity bit numbers with an estimated post-correction bit error rate not higher than the target post-correction bit error rate and a number of correctable symbols not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin further include: determining the RS symbol error probability based on the pre-correction bit error rate and the number of bits contained in each RS symbol; for each preliminary candidate RS coding parity bit number, estimating the codeword decoding failure probability based on the RS symbol error probability, the number of RS symbols contained in an RS codeword, and the number of correctable symbols in the preliminary candidate RS coding parity bit number; and determining the codeword decoding failure probability based on the codeword decoding failure probability. The probability and decoding failure residual error conversion factor are used to determine the estimated post-correction bit error rate of the preliminary candidate RS code check bits. The preliminary candidate RS code check bits with an estimated post-correction bit error rate not higher than the target post-correction bit error rate and a number of correctable symbols not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin are determined as reliable candidate RS code check bits, and a reliable candidate RS code check bit set is generated. When the reliable candidate RS code check bit set is empty, the output link error correction capability is insufficient, and the candidate RS code check bit with the largest number of correctable symbols in the candidate RS code check bit set is marked as the protection upper limit configuration.

[0015] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal RS code parity bit number described in this invention, the step of selecting the candidate RS code parity bit number with the lowest coding overhead as the optimal RS code parity bit number includes: forming a reliable candidate RS code parity bit number set from the filtered candidate RS code parity bit numbers, and determining the coding overhead of each candidate RS code parity bit number according to each candidate RS code parity bit number and the corresponding number of effective data bits; determining the candidate RS code parity bit number with the lowest coding overhead in the reliable candidate RS code parity bit number set as the optimal RS code parity bit number; when two or more candidate RS code parity bit numbers have the same lowest coding overhead, selecting the candidate RS code parity bit number with the lowest decoding delay; when the decoding delay is the same, selecting the candidate RS code parity bit number ranked first in the FEC configuration table; writing the optimal RS code parity bit number into the configuration registers of the RS forward error correction encoder and RS forward error correction decoder, and generating an FEC configuration identifier according to the optimal RS code parity bit number, RS symbol bit width, and RS code character number, so that the transmitting side and the receiving side complete the configuration according to the same optimal RS code parity bit number.

[0016] As a preferred embodiment of the forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS encoding described in this invention, the output of the error-corrected 1.6T PAM4 data stream includes: the transmitting-side RS forward error correction encoder dividing the 1.6T PAM4 data stream to be transmitted into several RS valid data blocks according to the effective data length; performing finite field operations on each RS valid data block to generate redundant parity symbols corresponding to the optimal parity bit length of RS encoding; and combining the RS valid data blocks and redundant parity symbols into RS codewords; sequentially performing serial-to-parallel conversion, channel allocation, scrambling processing, PAM4 bit mapping, and PAM4 level modulation on the RS codewords to generate a 1.6T PAM4 transmission signal and transmitting it to the high-speed communication link; and the receiving side processing the received 1.6T PAM4 signal. The PAM4 signal undergoes analog front-end sampling, clock data recovery, equalization processing, PAM4 level decision, and bit recovery to obtain the received bit stream. Based on the optimal parity bit length and RS codeword boundary information, the received bit stream is divided into corresponding RS codewords. The receiving-side RS forward error correction decoder performs synodal calculation on each RS codeword. When the synodal calculation result is zero, it outputs the corresponding valid RS data block. When the synodal calculation result is not zero, it performs error location and error amplitude calculation. If the number of error symbols does not exceed the number of error-correctable symbols corresponding to the optimal parity bit length of the RS codeword, it corrects the error symbols. The valid RS data blocks corresponding to the RS codewords that did not experience decoding failure are reassembled in their original order to obtain the error-corrected 1.6T PAM4 data stream.

[0017] The beneficial effects of this invention are as follows: it enables fine-grained perception of link errors through codeword-level error observation, supporting dynamic selection of the number of parity bits; it compensates for the lack of information in the average bit error rate by quantifying error aggregation indicators; it improves the success rate of error correction under poor channel conditions by aggregation correction margin compensation; and it reduces transmission overhead while ensuring reliability by selecting the minimum overhead parity bits. 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 a forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS encoding.

[0020] Figure 2 This is a flowchart for acquiring link observation data.

[0021] Figure 3A flowchart for determining the aggregation correction margin.

[0022] Figure 4 A flowchart for selecting the optimal parity bit length and forward error correction for RS encoding.

[0023] Figure 5 A comparison chart of bit error rates after error correction under different RS parity bit configurations.

[0024] Figure 6 A comparison chart of encoding overhead under different RS parity bit configurations. Detailed Implementation

[0025] 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.

[0026] 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.

[0027] 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.

[0028] Reference Figures 1-6 This is one embodiment of the present invention, which provides a forward error correction method for 1.6T PAM4 signals based on the optimal parity bit length of RS encoding, comprising the following steps:

[0029] S1. Collect link observation data of 1.6T PAM4 signal, establish RS codeword-level error observation window based on RS codeword boundary information, and obtain the number of error symbols, burst error coverage information and bit error rate before error correction for each RS codeword.

[0030] Furthermore, the 1.6T PAM4 signal is carried in parallel by multiple high-speed PAM4 physical channels. PAM4 level decision is performed on the received signal of each high-speed PAM4 physical channel to obtain the PAM4 decision symbol stream corresponding to each high-speed PAM4 physical channel.

[0031] Convert the PAM4 decision symbol stream into a bit data stream.

[0032] Based on the RS encoded frame structure, RS codeword length, RS symbol bit width, and codeword synchronization marker, RS codeword boundary identification is performed on the bit data stream to obtain RS codeword boundary information.

[0033] The RS codeword boundary information includes the codeword start position, codeword end position, codeword number, high-speed PAM4 physical channel number to which it belongs, and codeword reception timestamp for each RS codeword.

[0034] The boundary identification is based on the codeword synchronization mark or the fixed RS codeword length.

[0035] When a codeword synchronization marker is detected in the received bit data stream, the position of the first RS symbol after the codeword synchronization marker is determined as the starting position of the current RS codeword.

[0036] When the system uses a fixed RS codeword length, the RS codewords are divided according to the fixed RS codeword length order, with the starting position of the synchronized RS codeword as the reference.

[0037] Therefore, each RS codeword has a unique corresponding codeword number and codeword boundary position.

[0038] After obtaining the RS codeword boundary information, perform synergistic calculation, error location, and error correction for each RS codeword.

[0039] If the result of the syndrome calculation for a certain RS codeword is zero, it is determined that no error symbol was detected in the RS codeword during the current reception period.

[0040] If the result of the syndrome calculation of a certain RS codeword is not zero, the position of the error symbol in the RS codeword is determined according to the error location polynomial, and the number of error symbols in the RS codeword is counted.

[0041] Thus, the number of erroneous symbols for each RS codeword is obtained.

[0042] The number of error symbols in each RS codeword is counted based on the error flag of the RS symbol. That is, when the error flag of the RS symbol is 0, it means that the current RS symbol is not marked as an error; when the error flag of the RS symbol is 1, it means that the current RS symbol is marked as an error.

[0043] Furthermore, based on the RS codeword boundary information, multiple consecutively received RS codewords are divided into RS codeword-level error observation windows.

[0044] Preferably, an RS codeword-level error observation window includes consecutive RS codewords received by the same high-speed PAM4 physical channel within one observation period, in order to avoid inaccurate judgment of the degree of error clustering within the codeword due to the mixing of error characteristics from different channels.

[0045] Suppose that the current RS codeword-level error observation window includes There are 10 RS codewords, and the number of error symbols corresponding to each RS codeword forms a codeword-level error sequence.

[0046] In order to obtain burst error coverage information, based on the codeword-level error sequence and RS codeword boundary information, the time interval of consecutively occurring error symbols or the RS codeword interval of consecutively occurring non-zero error symbols is identified.

[0047] When the error interval between adjacent RS codewords or adjacent RS symbols is less than the burst interval threshold, the erroneous symbols are classified into the same burst error event.

[0048] When the error interval is greater than or equal to the burst interval threshold, the error symbols are divided into different burst error events.

[0049] For each burst error event, the number of RS codewords covered by the burst error event is determined based on the start position, end position and RS codeword boundary information of the burst error event, and the number of error symbols in each covered RS codeword is counted.

[0050] The burst error coverage information includes the burst error event number, the number of RS codewords covered by the burst error, and the number of error symbols in each covered RS codeword.

[0051] The number of RS codewords covered by the sudden bit error event is represented as follows: ;

[0052] In the formula, Indicates the first The number of RS codewords covered by a sudden bit error event; Indicates the first The start and end positions of each RS codeword; Indicates the first The start and end positions of a sudden bit error event; Indicates the intersection of intervals; Indicates the empty set; This indicates the number of elements in the set.

[0053] The number of error symbols in each covered RS codeword is obtained by counting the pollution markers of whether the RS symbols in the RS codeword fall into the coverage area of ​​the corresponding burst error event. That is, when the RS symbol falls into the coverage area of ​​the corresponding burst error event, the number of error symbols is increased by 1, otherwise it is 0.

[0054] Furthermore, the bit error rate before error correction is calculated based on the number of erroneous bits detected within the current RS codeword-level error observation window and the total number of received bits.

[0055] The bit error rate before error correction is expressed as: ;

[0056] in, This represents the bit error rate before correction within the current RS codeword-level error observation window. This indicates the number of error bits detected within the current RS codeword-level error observation window before RS error correction. This indicates the total number of bits received within the current RS codeword-level error observation window.

[0057] Furthermore, a link observation dataset corresponding to the same high-speed PAM4 physical channel is generated within the current observation period.

[0058] The link observation dataset includes RS codeword-level error observation windows, the number of error symbols for each RS codeword, burst error coverage information, and the bit error rate before error correction.

[0059] S2. Based on the number of erroneous symbols in each RS codeword, calculate the maximum number of erroneous symbols, the dispersion of the number of erroneous symbols, and the proportion of high-erroneous codewords to generate an error clustering index within the codeword.

[0060] Furthermore, the RS codeword-level error observation window and codeword-level error sequence are received, the codeword-level error sequence is traversed, and the maximum number of error symbols within the current RS codeword-level error observation window is obtained.

[0061] The maximum number of erroneous symbols is used to characterize the highest error correction pressure borne by a single RS codeword within the current observation window.

[0062] Furthermore, the average number of error symbols is calculated based on the codeword-level error sequence.

[0063] Among them, the average number of error symbols is used to characterize the average error level of each RS codeword within the current RS codeword-level error observation window.

[0064] Furthermore, the dispersion of the number of error symbols is calculated based on the mean number of error symbols.

[0065] Among them, the dispersion of the number of error symbols is used to characterize the fluctuation range of the number of error symbols between different RS codewords.

[0066] When the number of error symbols is highly discrete, it indicates that the error symbols are not evenly distributed across multiple RS codewords, but rather tend to be concentrated in a portion of the RS codewords.

[0067] The dispersion of the number of error symbols is represented by the standard deviation, which is calculated as follows: ;

[0068] in, This indicates the degree of dispersion in the number of error symbols within the current RS codeword-level error observation window. Indicates the first The number of error symbols within an RS codeword This represents the average number of error symbols within the current RS codeword-level error observation window. This indicates the number of RS codewords contained within the current RS codeword-level error observation window.

[0069] Furthermore, the proportion of high-error codewords is determined based on the error attention threshold.

[0070] Among them, the error concern threshold is used to distinguish between ordinary error codewords and high error codewords.

[0071] RS codewords with a number of error symbols greater than or equal to the error concern threshold are identified as high error codewords, and the number of high error codewords within the current RS codeword-level error observation window is counted.

[0072] Specifically, it is determined one by one whether the number of erroneous symbols in each RS codeword within the current RS codeword-level error observation window has reached the error concern threshold, and the number of RS codewords that have reached the error concern threshold is divided by the total number of RS codewords to obtain the proportion of high error codewords.

[0073] After obtaining the maximum number of error symbols, the dispersion of the number of error symbols, and the proportion of high error codewords, an error clustering index within the codeword is generated.

[0074] Among them, the error clustering index within a codeword is used to characterize whether error symbols are concentrated in a few RS codewords, rather than being evenly distributed among multiple RS codewords.

[0075] Specifically, the error clustering index within a codeword is calculated as follows: ;

[0076] in, This indicates the error clustering index within the codeword. This indicates the maximum number of error symbols within the current RS codeword-level error observation window. This represents the average number of error symbols within the current RS codeword-level error observation window. This indicates the degree of dispersion in the number of error symbols within the current RS codeword-level error observation window. This indicates the percentage of high-error codewords within the current RS codeword-level error observation window. , , These represent the adjustment coefficients for the maximum error deviation term, the dispersion term, and the high error codeword percentage term, respectively.

[0077] Among them, the adjustment coefficient , , Determined through link factory calibration, historical bit error statistics, or standard bit error injection, and meeting the following requirements. value range , , All are greater than or equal to 0.

[0078] It should be noted that the maximum error pressure a single RS codeword can withstand is identified by the maximum number of error symbols; the degree of uneven distribution of error symbols among multiple RS codewords is identified by the dispersion of the number of error symbols; and the proportion of RS codewords approaching the error correction limit is identified by the proportion of high-error codewords. This ultimately generates an error clustering index within the codeword. Therefore, it avoids judging the link error status solely based on the average pre-correction bit error rate, thus providing a codeword-level risk basis for determining the optimal parity bit length for RS encoding.

[0079] S3. Determine the clustering correction margin based on the error clustering index within the codeword and the burst error coverage information.

[0080] Furthermore, based on the error clustering index within the codeword, it is determined whether the error symbols within the current RS codeword-level error observation window have a local concentration trend, and based on the burst error coverage information, it is determined whether burst errors mainly fall into a few RS codewords.

[0081] It should be noted that if the error symbols are concentrated in a few RS codewords, or if a sudden error causes a large number of error symbols in a single or a few RS codewords, it indicates that there is a risk of local decoding failure in the current RS codeword-level error observation window, and a clustering correction margin needs to be added during the candidate RS code check bit selection process.

[0082] Determine the clustering correction margin; specifically, calculate the burst error concentration based on burst error coverage information.

[0083] For each burst error event within the current RS codeword-level error observation window, count the number of RS codewords covered by the burst error event and the number of error symbols within each covered RS codeword.

[0084] It should be noted that if the number of RS codewords covered by a burst error is small, but the number of error symbols in a single covered RS codeword is high, it indicates that the concentration of burst errors is high.

[0085] The formula for calculating the burst error concentration is: ;

[0086] in, This indicates the concentration of burst errors within the current RS codeword-level error observation window. Indicates the sudden error event number, Indicates the first The first sudden error event was in the The number of error symbols caused within an RS codeword This indicates the number of RS codewords contained within the current RS codeword-level error observation window. Indicates the first The number of RS codewords covered by a sudden bit error event. This indicates taking the maximum value among all burst error events within the current RS codeword-level error observation window.

[0087] Furthermore, the error clustering index within the codeword and the concentration of sudden errors are combined to obtain the clustering risk value.

[0088] Among them, the cluster risk value is used to uniformly characterize the risk of underestimation of the parity bit depth caused by the concentrated distribution of erroneous symbol codewords and the local contamination of sudden errors.

[0089] Cluster risk value, its calculation formula is as follows: ;

[0090] in, Indicates the cluster risk value. This indicates the error clustering index within the codeword. Indicates the concentration of sudden errors. This represents the adjustment coefficient corresponding to the error clustering index within the codeword. This represents the adjustment coefficient corresponding to the concentration of sudden bit errors.

[0091] After obtaining the cluster risk value, the cluster correction margin is determined based on the cluster risk level at which the cluster risk value is located.

[0092] Among them, the clustering correction margin represents the additional number of error-correctable symbols required to cover the risk of concentrated error symbols within a local RS codeword during the screening process of candidate RS code check bits.

[0093] The cluster risk level segmentation rules for cluster adjustment margin are as follows:

[0094] When the cluster risk value is less than the first cluster risk threshold, the cluster risk of error symbols within the current RS codeword-level observation window is no risk, and no cluster correction margin is added.

[0095] When the cluster risk value is not less than the first cluster risk threshold and the cluster risk value is less than the second cluster risk threshold, the cluster risk of erroneous symbols in the current RS codeword-level observation window is low, and then the cluster correction margin of 1 RS symbol is increased.

[0096] When the cluster risk value is not less than the second cluster risk threshold and the cluster risk value is less than the third cluster risk threshold, the cluster risk of erroneous symbols in the current RS codeword-level observation window is medium risk, and then the cluster correction margin of 2 RS symbols is increased.

[0097] When the cluster risk value is not less than the third cluster risk threshold, the cluster risk of erroneous symbols in the current RS codeword-level observation window is high, and then the cluster correction margin of 3 RS symbols is increased.

[0098] The first, second, and third clustering risk thresholds can be determined through factory error injection experiments. Specifically, in a 1.6T PAM4 link test platform, random errors, short burst errors, and long burst errors are injected respectively, and the distribution of clustering risk values ​​before RS decoding failure occurs under different error clustering states is recorded. The upper limit of the clustering risk value when no decoding failure occurs is set as the first clustering risk threshold, the clustering risk value when the decoding failure probability reaches a first preset proportion is set as the second clustering risk threshold, and the clustering risk value when the decoding failure probability reaches a second preset proportion is set as the third clustering risk threshold. The first preset proportion can be 1% to 5%, and the second preset proportion can be 5% to 10%.

[0099] It should be noted that the aggregation correction margin is used for screening the check bits of candidate RS codes.

[0100] Specifically, when selecting candidate RS code check bits, the number of error-correctable symbols corresponding to the candidate RS code check bits needs to cover not only the maximum number of error symbols within the current RS codeword-level error observation window, but also the aggregation correction margin, i.e., satisfy... .

[0101] in, Indicates the first The number of error-correctable symbols corresponding to each candidate RS code check bit. This indicates the maximum number of error symbols within the current RS codeword-level error observation window. This indicates the aggregated correction margin.

[0102] S4. Based on the target bit error rate after error correction, the bit error rate before error correction, the set of candidate RS code check bits, the maximum number of erroneous symbols, and the aggregation correction margin, select candidate RS code check bits whose estimated bit error rate after error correction is not higher than the target bit error rate after error correction, and whose number of correctable symbols is not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin.

[0103] Furthermore, the candidate RS code check bits are selected based on the target bit error rate after error correction, the bit error rate before error correction, the set of candidate RS code check bits, the maximum number of erroneous symbols, and the aggregation correction margin.

[0104] The target bit error rate after error correction is 1.6T. The maximum bit error rate allowed by the PAM4 link under the system reliability requirements is pre-written by the communication protocol, link management controller, optical module management interface or physical layer configuration register of the switching chip.

[0105] The bit error rate before correction is the statistical result of the bit error rate before RS error correction within the current RS codeword-level error observation window.

[0106] The candidate RS code check bit set consists of multiple RS code check bit configurations, which can be pre-stored in a register group, on-chip static random access memory, or FEC configuration table.

[0107] The number of check bits for each candidate RS code is converted into the number of error-correctable symbols in the candidate RS code.

[0108] If each RS symbol in the RS encoding contains Then the bit The number of redundant check symbols corresponding to the number of candidate RS code check bits is: ;

[0109] in, Indicates the first The number of redundant check symbols corresponding to each candidate RS code check bit. Indicates the first One candidate RS code check bit, This indicates the number of bits contained in each RS symbol.

[0110] For RS encoding, if it has If there are redundant check symbols, then the number of error-correcting symbols is: ;

[0111] in, Indicates the first The number of error-correctable symbols corresponding to each candidate RS code check bit. Indicates the first The number of redundant check symbols corresponding to the number of candidate RS code check bits.

[0112] Furthermore, based on the maximum number of erroneous symbols and the clustering correction margin, the minimum error-correctable symbol requirement corresponding to the current RS codeword-level error observation window is determined.

[0113] The minimum error-correctable symbol requirement represents the codeword-level error risk that the current candidate RS code check bits must at least cover, and it is calculated as follows: ;

[0114] in, This represents the minimum required number of error-correctable symbols. This indicates the maximum number of error symbols within the current RS codeword-level error observation window. This indicates the aggregated correction margin.

[0115] Subsequently, the first screening condition judgment is performed on each candidate RS code check bit.

[0116] If the first The number of error-correctable symbols corresponding to the number of candidate RS code check bits satisfies Then the first The number of candidate RS code check bits is selected based on codeword-level error correction capability.

[0117] If the first If the number of error-correctable symbols corresponding to the number of candidate RS code check bits is less than the minimum error-correctable symbol requirement, then the candidate RS code check bit will be eliminated. One candidate RS code check bit.

[0118] Furthermore, based on the bit error rate before error correction and the number of error-correctable symbols in the candidate RS code, the estimated bit error rate after error correction corresponding to the parity bits of each candidate RS code is calculated.

[0119] Specifically, the bit error rate before error correction is converted into the RS symbol error probability: ;

[0120] in, Indicates the RS sign error probability. This represents the bit error rate before error correction. This indicates the number of bits contained in each RS symbol.

[0121] Then, based on the RS symbol error probability and the number of error-correctable symbols in the candidate RS codes, the number of symbols is estimated. The probability of codeword decoding failure corresponding to each candidate RS code check bit: ;

[0122] in, Indicates the first The probability of codeword decoding failure corresponding to each candidate RS code check bit. Indicates the first The number of error-correctable symbols corresponding to each candidate RS code check bit. This indicates the number of RS symbols contained in an RS codeword. This indicates the number of RS symbols that cause errors within an RS codeword. Indicates from Select from RS symbols The number of combinations of erroneous RS symbols, This represents the probability of RS sign error.

[0123] Furthermore, the post-correction bit error rate is estimated based on the codeword decoding failure probability: ;

[0124] in, Indicates the first The estimated bit error rate after error correction corresponds to the number of candidate RS code check bits. Indicates the first The probability of codeword decoding failure corresponding to each candidate RS code check bit. Indicates the first The residual error conversion factor for decoding failure corresponding to each candidate RS code check bit.

[0125] In obtaining the first After estimating the bit error rate after error correction for each candidate RS code check bit, the check bit selection unit performs the second selection condition judgment.

[0126] If satisfied Then the first The candidate RS code check bits are filtered by the bit error rate after error correction; if it does not meet the requirements... Then remove the first one. One candidate RS code check bit.

[0127] in, Indicates the first The estimated bit error rate after error correction corresponds to the number of candidate RS code check bits. This represents the bit error rate after the target error correction.

[0128] The candidate RS code check bits that simultaneously meet the first and second screening conditions are determined as reliable candidate RS code check bits, and a set of reliable candidate RS code check bits is generated.

[0129] If the set of reliable candidate RS code check bits is empty, the output link error correction capability is insufficient, and the candidate RS code check bit with the largest number of error-correctable symbols in the candidate RS code check bit set is marked as the protection upper limit configuration, so that the link management controller can perform speed reduction, retraining, alarm or optical power adjustment.

[0130] It should be noted that when the set of reliable candidate RS coding check bits is empty, the step of selecting the candidate RS coding check bits with the lowest coding cost is not executed. Instead, the protection upper limit configuration is used as the temporary RS forward error correction configuration, and a link retraining or speed reduction prompt is triggered. After regenerating the set of reliable candidate RS coding check bits in the next observation period, the optimal RS coding check bits are selected based on the coding cost.

[0131] S5. Among the candidate RS code check bits after screening, select the candidate RS code check bits with the lowest coding overhead as the RS code optimal check bits, and perform RS forward error correction coding and decoding on the 1.6T PAM4 data stream according to the RS code optimal check bits, and output the error-corrected 1.6T PAM4 data stream.

[0132] Furthermore, a set of reliable candidate RS code check bits is received.

[0133] Based on the set of reliable candidate RS code check bits, calculate the coding cost corresponding to each candidate RS code check bit in the set of reliable candidate RS code check bits.

[0134] The encoding overhead is used to characterize the proportion of redundant check data in the RS codeword relative to the total data in the codeword.

[0135] Specifically, no. The coding overhead corresponding to each candidate RS code parity bit is calculated using the following formula: ;

[0136] in, Indicates the first The coding overhead corresponding to the number of candidate RS code check bits. Indicates the first One candidate RS code check bit, Indicates the first The number of valid data bits corresponding to each candidate RS code check bit.

[0137] After obtaining the coding overhead corresponding to each reliable candidate RS code parity bit, select the candidate RS code parity bit with the lowest coding overhead from the set of reliable candidate RS code parity bit bits as the optimal RS code parity bit. ;

[0138] in, Indicates the optimal parity bit length for RS encoding. Represents the set of reliable candidate RS code check bits. Represents the first bit in the set of reliable candidate RS code check bits. One candidate RS code check bit, Indicates the first The coding overhead corresponding to the number of candidate RS code check bits. This indicates the selection of the candidate RS code check bit that minimizes the coding overhead from the set of reliable candidate RS code check bit bits.

[0139] If two or more candidate RS code check bits have the same minimum coding overhead, the candidate RS code check bit with the lower corresponding decoding delay is selected first; if the decoding delay is also the same, the candidate RS code check bit with the higher ranking in the FEC configuration table is selected.

[0140] After determining the optimal parity bit length for RS encoding, the optimal parity bit length for RS encoding is written into the configuration registers of the RS forward error correction encoder and the RS forward error correction decoder.

[0141] To ensure that the sending and receiving sides use the same RS-encoded optimal parity bit, the RS-encoded optimal parity bit is synchronously written into the encoding and decoding paths. For scenarios where the two ends of the link are set independently, the sending side sends the FEC configuration identifier corresponding to the RS-encoded optimal parity bit to the receiving side through the link training sequence, management channel, physical layer control field, or FEC configuration negotiation field.

[0142] After confirming the FEC configuration identifier, the receiving side configures the RS forward error correction decoder according to the same RS encoding optimal parity bit number.

[0143] The FEC configuration identifier is generated based on the optimal parity bit length and RS symbol bit width of the RS encoding. It maps the optimal parity bit length, RS symbol bit width, and RS code character number to a unique FEC configuration identifier, enabling the sending and receiving sides to confirm the currently used RS forward error correction configuration through the FEC configuration identifier.

[0144] After completing the optimal parity bit configuration for RS encoding, the transmitting-side RS forward error correction encoder performs RS encoding on the 1.6T PAM4 data stream to be transmitted according to the optimal parity bit configuration for RS encoding.

[0145] Specifically, the transmitting-side RS forward error correction encoder divides the 1.6T PAM4 data stream to be transmitted into multiple RS valid data blocks according to the effective data length; performs finite field operations on each RS valid data block to generate a redundant check symbol corresponding to the optimal check bit length of the RS encoding; and combines the RS valid data blocks and the redundant check symbols to generate RS codewords.

[0146] Furthermore, the RS codeword is subjected to serial-to-parallel conversion, channel allocation, scrambling, PAM4 bit mapping, and PAM4 level modulation to generate a 1.6T PAM4 transmit signal and send it to the high-speed communication link.

[0147] For multi-channel 1.6T PAM4 links, the transmitting side maps RS codewords to the corresponding high-speed PAM4 physical channels according to the channel allocation rules, so that each high-speed PAM4 physical channel adopts the optimal parity bit of RS encoding corresponding to the current FEC configuration identifier.

[0148] The received 1.6T PAM4 signal is subjected to analog front-end sampling, clock data recovery, equalization processing, PAM4 level decision and bit recovery to obtain the received bit stream.

[0149] The receiving-side RS forward error correction decoder divides the received bit stream into corresponding RS codewords based on the optimal parity bit length of the RS encoding and the RS codeword boundary information, and performs synodal calculations on each RS codeword.

[0150] If the result of the synergy calculation is zero, the receiver-side RS forward error correction decoder determines that no error has been detected in the current RS codeword and directly outputs the corresponding valid data block; if the result of the synergy calculation is not zero, the receiver-side RS forward error correction decoder performs error location and error magnitude calculation according to the number of error-correctable symbols corresponding to the optimal parity bits of the RS encoding, and corrects the erroneous symbols.

[0151] The receiver-side RS forward error correction decoder determines the number of error-correctable symbols based on the optimal parity bit depth of the RS encoding. ;

[0152] in, This indicates the number of error-correctable symbols corresponding to the optimal check bit length in RS encoding. Indicates the optimal parity bit length for RS encoding. This indicates the number of bits contained in each RS symbol.

[0153] The number of error symbols corresponding to the error location result of the receiver-side RS forward error correction decoder does not exceed When the error location result corresponds to more than 100 error symbols, perform error correction and output the corrected RS valid data block; when the number of error symbols corresponds to more than 100 error symbols, perform error correction and output the corrected RS valid data block. When the decoding fails, output a decoding failure flag and mark the corresponding codeword as an uncorrectable codeword.

[0154] It should be noted that for RS codewords that fail to decode, the receiving-side RS forward error correction decoder outputs a decoding failure flag and records the RS codeword in the next RS codeword-level error observation window. Here, the decoded codeword is not used as a codeword that can accurately locate the number of error symbols in the error symbol count, but is used as an unreliable statistical codeword in the judgment of high error codeword ratio, burst error coverage information and insufficient link error correction capability.

[0155] For RS codewords that do not experience decoding failure, the receiver-side RS forward error correction decoder reassembles the error-corrected RS valid data blocks in their original order to obtain the error-corrected 1.6T PAM4 data stream.

[0156] In this embodiment, Figure 5The results show a comparison of the bit error rate after error correction as a function of the RS codeword-level error observation window under different RS parity bit configurations.

[0157] Depend on Figure 5 As can be seen, within the observation window where the link operates relatively smoothly, the optimal adaptive parity bit configuration, the fixed 128-bit parity bit configuration, and the fixed 192-bit parity bit configuration adopted in this invention can all maintain the post-correction bit error rate at a low level. When burst errors and local clustering of erroneous symbols occur in observation windows 18–21, 39–43, and 59–61, the post-correction bit error rate of the fixed 128-bit parity bit configuration increases significantly, exceeding the target post-correction bit error rate of 1×10^-12 multiple times; although the fixed 192-bit parity bit configuration can maintain a low error rate, it always uses a high redundancy configuration. This invention determines the clustering correction margin based on the maximum number of erroneous symbols, the error clustering index within the codeword, and the burst error coverage information, and promptly increases the RS coding parity bit, so that the post-correction bit error rate remains below the target value during burst interference.

[0158] This demonstrates that the present invention can identify the risk of local error clustering that is difficult to reflect by the average bit error rate before error correction, reduce the probability of RS codeword decoding failure, and improve the transmission reliability and anti-sudden interference capability of the 1.6T PAM4 link under dynamic bit error conditions.

[0159] In this embodiment, Figure 6 The results show a comparison of encoding overhead as a function of the RS codeword-level error observation window under different RS parity bit configurations.

[0160] Depend on Figure 6 It is evident that the coding overhead of a fixed 128-bit parity configuration remains consistently around 5.9%, and the coding overhead of a fixed 192-bit parity configuration remains consistently around 8.6%. Neither of these fixed configurations can adjust redundant resources according to the actual error state of the link. This invention employs an adaptive configuration of the optimal parity bit count. Within most observation windows where the link error risk is low, the coding overhead is maintained at approximately 4.5%. When a higher error clustering risk occurs in observation windows 18–21, 39–43, and 59–61, the coding overhead increases sequentially to approximately 5.9%, 7.2%, or 8.6%, and then returns to a lower level after the error risk decreases. This variation indicates that this invention does not continuously use the maximum parity bit count. Instead, it selects the candidate RS coding parity bit count with the lowest coding overhead as the optimal RS coding parity bit count, based on the fact that the candidate RS coding parity bit count simultaneously meets the requirements for the number of error-correctable symbols and the target post-correction bit error rate.

[0161] This demonstrates that the present invention can reduce unnecessary redundant verification data while ensuring the reliability of link error correction, increase the effective data transmission ratio in the 1.6T PAM4 data stream, and achieve a dynamic balance between error correction capability and coding overhead.

[0162] In summary, this invention achieves refined perception of link bit errors through codeword-level error observation, supporting dynamic selection of check bits; it compensates for the lack of information in the average bit error rate by quantifying error aggregation indicators; it improves the error correction success rate under poor channel conditions by aggregation correction margin compensation; and it reduces transmission overhead while ensuring reliability by selecting the minimum overhead check bits.

[0163] 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 forward error correction method for 1.6T PAM4 signals based on the optimal parity bit depth of RS encoding, characterized in that, include: Link observation data of 1.6T PAM4 signal were collected, and RS codeword-level error observation window was established based on RS codeword boundary information to obtain the number of error symbols, burst error coverage information and bit error rate before error correction for each RS codeword; Based on the number of erroneous symbols in each RS codeword, calculate the maximum number of erroneous symbols, the dispersion of the number of erroneous symbols, and the proportion of high-erroneous codewords to generate an error clustering index within the codeword; The clustering correction margin is determined based on the error clustering index within the codeword and the burst error coverage information. Based on the target bit error rate after error correction, the bit error rate before error correction, the set of candidate RS code check bits, the maximum number of erroneous symbols, and the aggregation correction margin, candidate RS code check bits are selected that are not higher than the target bit error rate after error correction and the number of erroneous symbols that can be corrected is not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin. Among the candidate RS code check bits selected after screening, the candidate RS code check bits with the lowest coding overhead are selected as the RS code optimal check bits. RS forward error correction coding and decoding are performed on the 1.6T PAM4 data stream according to the RS code optimal check bits, and the error-corrected 1.6T PAM4 data stream is output. The burst error coverage information includes the burst error event number, the number of RS codewords covered by the burst error, and the number of error symbols in each covered RS codeword; The error clustering index within the codeword is used to characterize whether error symbols are concentrated in a small number of RS codewords; The aggregation correction margin represents the additional number of error-correctable symbols required to cover the risk of concentrated error symbols within a local RS codeword during the candidate RS code check bit selection process.

2. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 1, characterized in that, The process of obtaining the number of erroneous symbols, burst error coverage information, and pre-correction error rate for each RS codeword includes: PAM4 level determination is performed on the received signals of each high-speed PAM4 physical channel; Convert the obtained PAM4 decision symbol stream into a bit data stream; Determine the starting position of the RS codeword based on the codeword synchronization marker in the bit data stream; Based on a determined starting position, RS codewords are divided according to their length order to obtain RS codeword boundary information; Perform a synodal computation for each RS codeword; When the result of the adjoint calculation is not zero, the position of the error sign is determined based on the error location polynomial; The number of error symbols in each RS codeword is counted based on the error flags of the RS symbols. The RS codewords continuously received by the same high-speed PAM4 physical channel within one observation period are divided into RS codeword-level error observation windows. Adjacent error symbols with an error interval less than the burst interval threshold are grouped into the same burst error event; Based on the mapping relationship between burst error events and RS codeword boundaries, the number of RS codewords covered by burst error events and the number of error symbols in each covered RS codeword are counted to obtain burst error coverage information. The bit error rate before correction is calculated based on the ratio of the number of erroneous bits detected before RS error correction within the RS codeword-level error observation window to the total number of received bits.

3. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 1, characterized in that, The specific steps for calculating the maximum number of error symbols, the dispersion of error symbol counts, and the proportion of high-error codewords based on the number of error symbols in each RS codeword are as follows: Traverse the codeword-level error sequence corresponding to the current RS codeword-level error observation window; The maximum value in the codeword-level error sequence is determined as the maximum number of error symbols; Calculate the average number of error symbols for each RS codeword within the current RS codeword-level error observation window, and determine the degree of dispersion of the number of error symbols based on the degree of deviation of the number of error symbols for each RS codeword from the average number of error symbols. RS codewords with a number of error symbols greater than or equal to the error concern threshold are identified as high error codewords; The percentage of high-error codewords is determined by the proportion of the number of high-error codewords to the total number of RS codewords within the current RS codeword-level error observation window.

4. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 3, characterized in that, The generated codeword error clustering index includes: Calculate the difference between the maximum number of error symbols and the mean number of error symbols, and normalize the difference using the mean number of error symbols to obtain the maximum error deviation term; The dispersion of the number of error symbols is normalized using the mean of the number of error symbols to obtain the dispersion term; The maximum error deviation term, the dispersion term, and the high error codeword ratio term are adjusted according to their respective adjustment coefficients. The adjusted maximum error deviation term, dispersion term, and high error codeword ratio term are then superimposed to obtain the error clustering index within the codeword.

5. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 1, characterized in that, The determination of the aggregation correction margin includes: For each burst error event within the current RS codeword-level error observation window, count the number of RS codewords covered by the burst error event and the number of error symbols within each covered RS codeword; Based on the maximum number of erroneous symbols, the total number of erroneous symbols, and the number of covered RS codewords for each burst error event, the burst error concentration corresponding to each burst error event is determined, and the maximum value among the burst error concentrations corresponding to all burst error events is determined as the burst error concentration of the current RS codeword-level error observation window. The error clustering index and the burst error concentration within the codeword are adjusted according to the corresponding adjustment coefficients, and the adjusted error clustering index and burst error concentration are superimposed to obtain the clustering risk value.

6. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 5, characterized in that, The determination of the aggregation correction margin also includes: The clustering risk value is compared with the first clustering risk threshold, the second clustering risk threshold, and the third clustering risk threshold, respectively. When the clustering risk value is less than the first clustering risk threshold, the clustering correction margin is set to zero RS symbols; When the clustering risk value is not less than the first clustering risk threshold and less than the second clustering risk threshold, the clustering correction margin is determined as an RS symbol. When the clustering risk value is not less than the second clustering risk threshold and is less than the third clustering risk threshold, the clustering correction margin is determined as two RS symbols. When the clustering risk value is not less than the third clustering risk threshold, the clustering correction margin is determined as three RS symbols.

7. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 1, characterized in that, The specific steps for selecting candidate RS code check bits whose estimated post-correction bit error rate is not higher than the target post-correction bit error rate and whose number of correctable symbols is not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin are as follows: Based on the number of bits contained in each RS symbol, the number of bits of each candidate RS code check bit is converted into the number of redundant check symbols, and the number of error-correctable symbols is determined based on the number of redundant check symbols. The minimum number of correctable symbols is determined by summing the maximum number of erroneous symbols and the aggregation correction margin. Candidate RS code check bits with a number of error-correctable symbols not less than the minimum error-correctable symbol requirement are determined as preliminary candidate RS code check bits, and candidate RS code check bits with a number of error-correctable symbols less than the minimum error-correctable symbol requirement are eliminated.

8. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 7, characterized in that, The steps for selecting candidate RS codes with a post-correction bit error rate not higher than the target post-correction bit error rate and a number of correctable symbols not less than the sum of the maximum number of erroneous symbols and the aggregation correction margin are as follows: The error probability of an RS symbol is determined based on the bit error rate before error correction and the number of bits contained in each RS symbol. For each preliminary candidate RS code check bit, the codeword decoding failure probability is estimated based on the RS symbol error probability, the number of RS symbols contained in an RS codeword, and the number of error-correctable symbols in the preliminary candidate RS code check bit. Based on the codeword decoding failure probability and the residual error conversion factor of decoding failure, the estimated error rate after error correction is determined for the number of check bits of the preliminary candidate RS code. The preliminary candidate RS code check bits that are not higher than the target error rate after error correction and whose number of correctable symbols is not less than the sum of the maximum number of error symbols and the aggregation correction margin are determined as reliable candidate RS code check bits, and a set of reliable candidate RS code check bits is generated. When the set of reliable candidate RS code check bits is empty, the output link error correction capability is insufficient, and the candidate RS code check bit with the largest number of error-correctable symbols in the candidate RS code check bit set is marked as the protection upper limit configuration.

9. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 1, characterized in that, The selection of the candidate RS code parity bit with the lowest coding overhead as the optimal RS code parity bit includes: The selected candidate RS code check bits are combined into a reliable candidate RS code check bit set, and the coding overhead of each candidate RS code check bit is determined based on each candidate RS code check bit and the corresponding number of valid data bits. In the set of reliable candidate RS code parity bits, the candidate RS code parity bits with the lowest coding overhead are determined as the optimal RS code parity bits; When two or more candidate RS code parity bits have the same minimum coding overhead, the candidate RS code parity bit with the lowest decoding delay is selected. When the decoding delay is the same, select the candidate RS code check bit number that is ranked first in the FEC configuration table; Write the optimal parity bit length of the RS encoding into the configuration registers of the RS forward error correction encoder and RS forward error correction decoder, and generate an FEC configuration identifier based on the optimal parity bit length of the RS encoding, the RS symbol bit width, and the number of RS code characters, so that the transmitting and receiving sides complete the configuration with the same optimal parity bit length of the RS encoding.

10. The forward error correction method for 1.6T PAM4 signals based on the optimal parity bit of RS encoding as described in claim 9, characterized in that, The output of the 1.6T PAM4 data stream after error correction includes: The transmitting-side RS forward error correction encoder divides the 1.6T PAM4 data stream to be transmitted into several RS valid data blocks according to the effective data length. It performs finite field operations on each RS valid data block to generate redundant check symbols corresponding to the optimal check bits of RS encoding, and combines the RS valid data blocks and redundant check symbols into RS codewords. The RS codeword is sequentially processed by serial-to-parallel conversion, channel allocation, scrambling, PAM4 bit mapping, and PAM4 level modulation to generate a 1.6T PAM4 transmit signal and send it to the high-speed communication link. The receiving side performs analog front-end sampling, clock data recovery, equalization processing, PAM4 level decision and bit recovery on the received 1.6T PAM4 signal to obtain the received bit stream, and divides the received bit stream into corresponding RS codewords according to the optimal parity bit of RS encoding and RS codeword boundary information. The receiver-side RS forward error correction decoder performs a synodal calculation on each RS codeword. When the synodal calculation result is zero, it outputs the corresponding valid RS data block. When the synodal calculation result is not zero, it performs error location and error magnitude calculation. When the number of error symbols does not exceed the number of error-correctable symbols corresponding to the optimal parity bit of RS encoding, it corrects the error symbols. The valid RS data blocks corresponding to the RS codewords that did not fail to decode are reassembled in their original order to obtain the 1.6T PAM4 data stream after error correction.

Images