Improved TPC error correction algorithm and device
By iteratively correcting the uncertain symbols of single row or single column in TPC decoding technology, the problem of bottleneck in error correction ability in TPC decoding technology is solved, and higher error correction ability and higher decoding efficiency are achieved , suitable for application scenarios with high channel stability.
Patent Information
- Application Number
- CN202010885717.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-08-28
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2040-08-28
AI Technical Summary
There are bottlenecks in the error correction capability of existing TPC decoding technologies, especially when the number of row or column error symbols exceeds 2X+1 or 2Y+1.
By iteratively correcting errors in the error symbol matrix, inverse operations are performed on uncertain symbols in a single row or single column until the error correction is successful, the error correction ability is improved, and row and column correction is improved without increasing the code length. Wrong ability.
The error correction of all symbols is achieved within a limited number of iterations, which improves the row error correction ability to 2X+3 and the column error correction ability to 2Y+3, which reduces the hardware resource requirements, improves the decoding efficiency, and saves Storage space.
Smart Images

Figure CN112054809B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of TPC decoding, and in particular to an improved TPC error correction algorithm and device. Background Art
[0002] Turbo product code, referred to as TPC, has excellent error correction capability and low hardware implementation complexity. It has attracted great attention in the coding field and is also a popular coding technology. TPC forms a multi-dimensional error correction matrix during encoding, and then performs error correction on different dimensions during decoding. Due to the intersection of dimensions, there are multiple correction opportunities for a single code element, so the error correction capability is significantly improved. TPC is usually used for channel error correction coding interfered by burst noise. From the perspective of coding theory, almost all codes are asymptotically good codes without limiting the code length. However, due to the increase in code length, the transmission performance will be greatly affected, and the storage medium also has limitations on the storage space of the error correction code. Therefore, under the premise of balancing the error correction capability, storage capacity, and transmission performance, the length of the error correction code in each dimension is limited, and its error correction capability is also limited. Taking two-dimensional TPC as an example, the maximum row error correction capability is X code elements, and the maximum column error correction capability is Y code elements. That is, when the row error codeword is greater than X, the row error correction fails; when the column error codeword is greater than Y, the column error correction fails; when the row error codeword intersects with the column error correction codeword, the codeword cannot be corrected. In traditional TPC decoding, multiple iterative decoding or parallel decoding with different decoding orders and then optimal decoding are performed, which can improve the gain to a certain extent. However, if the same error correction failure rows and columns appear in subsequent iterative decoding, then the traditional TPC decoding has reached the error correction bottleneck.
[0003] In order to improve the error correction capability, the prior art patent applications CN110896309A (Decoding method, device, decoder and computer storage medium of Turbo product code) and CN111277355A (A method and device for correcting deadlock in TPC decoding) add an error correction process after decoding, that is, by inverting the code elements in the row and column matrix where the error correction fails, so as to break the error matrix deadlock phenomenon that occurs during the TPC decoding process. Although the scheme of inverting the code elements in the row and column matrix where the error correction fails breaks the original error structure, it can only effectively improve the error correction of the corresponding row or column when the number of row code element errors is less than or equal to X or the number of column code element errors is less than or equal to Y after inversion. In other words, the row error correction capability can be increased to 2X+1 code elements at most, and the column error correction capability can be increased to 2Y+1 code elements. When the number of row error code elements is greater than 2X+1 and the number of column error code elements is greater than 2Y+1, it is impossible to correct all code elements.
[0004] Therefore, in view of the defects of the prior art, it is necessary to propose a technical solution to solve the technical problems existing in the prior art. Summary of the invention
[0005] In view of this, it is indeed necessary to provide an improved TPC error correction algorithm and device, which, based on locating the uncorrectable error codeword matrix, further performs iterative error correction on the single row / column codewords in the error codeword matrix, so that the error matrix can quickly converge to the correctable range, thereby improving the error correction capability of TPC.
[0006] In order to solve the technical problems existing in the prior art, the technical solution of the present invention is as follows:
[0007] An improved TPC error correction algorithm includes at least the following steps:
[0008] Step S1: Obtain coded data and execute TPC decoding algorithm to generate an error symbol matrix;
[0009] Step S2: randomly selecting a row or column of data in the error symbol matrix;
[0010] Step S3: modify some code elements of the row or column, perform decoding operation on the row or column until decoding is successful, and update the error code element matrix;
[0011] Step S4: Repeat the above steps until there is no erroneous data in the error symbol matrix and output the decoded data.
[0012] As a further improvement, in step S3, the code elements in the corresponding row or column may be modified according to the number of uncertain code elements in the row or column; assuming that the number of uncertain code elements is K, the following steps are performed:
[0013] Step S31: Select K+1 symbols in the selected row or column according to the position information in the error matrix;
[0014] Step S32: performing a negation operation on the selected K+1 code elements;
[0015] Step S33: Execute the TPC decoding algorithm on the row or column after the codeword is modified. If the decoding is unsuccessful, perform all inversion operations on the selected row or column according to the position information in the error matrix and then execute the TPC decoding algorithm again. If the decoding is still unsuccessful, re-execute step S31 until all iterative processes are completed; if the decoding is successful, execute step S4.
[0016] As a further improvement, when the uncertain symbol K in a single row or column in the error matrix is 0, the step S31 is:
[0017] According to the position information of the error matrix, one code element of the selected row or column is inverted;
[0018] The error correction of the row or column can be successfully performed by executing step S33 again.
[0019] As a further improvement, when the uncertain code element K in a single row or column in the error matrix is 1, the specific steps are as follows:
[0020] When step S31 is executed for the first time, two code elements are optionally inverted in the selected row or column according to the position information of the error matrix. When the two code elements are of the same nature, step S33 can be executed to successfully correct the error; when the two code elements are of different natures, the first execution of step S33 fails and step S31 is executed again.
[0021] When executing step S31 for the second time, select one of the code elements selected for the first time, and then select another code element according to the position of the error matrix to invert them together. When the two code elements selected for the second time are code elements of the same nature, the second step S33 can successfully correct the error; when the two code elements are code elements of different natures, the second execution of step S33 fails. At this time, it can be determined that the other code element selected for the first time and the newly selected code element for the second time are code elements of the same nature. When executing step S31 for the third time, invert these two code elements together, which can ensure that the third step S33 is successfully corrected.
[0022] As a further improvement, step S1 includes performing a forward decoding operation, which is a traditional TPC decoding process. If the decoding is successful, the decoding result is directly output; otherwise, the original codeword and the error codeword matrix are output.
[0023] As a further improvement, step S1 further includes a negative decoding operation, which performs the following steps:
[0024] Obtain the error code element matrix output after the positive decoding operation;
[0025] According to the coordinate information of the error symbol matrix, the error rows or columns are inverted in sequence and a decoding operation is performed;
[0026] If a row or a column is decoded successfully, the decoding result is output, and the error code element matrix is updated and output at the same time; otherwise, the original code element and the error code element matrix are output.
[0027] As a further improvement, in step S2, the row or column with the least number of uncertain code elements is selected.
[0028] As a further improvement, in step S4, steps S2 to S3 are repeatedly executed until all rows or columns are decoded successfully.
[0029] As a further improvement, in step S4, steps S2 to S3 are repeatedly executed to make the error symbol matrix within the error correction capability range of the TPC decoding algorithm, and then step S1 is executed.
[0030] The present invention also discloses an improved TPC error correction device, which at least comprises:
[0031] A first acquisition unit, used to acquire coded data;
[0032] A generating unit, configured to execute a TPC decoding algorithm on the coded data to generate an error symbol matrix;
[0033] A second acquisition unit is used to acquire data of any selected row or column in the error symbol matrix;
[0034] A decoding unit, used for modifying part of the code elements of the selected row or column, and performing a decoding operation on the row or column until the decoding is successful to update the error code element matrix;
[0035] A judging unit, used to judge whether the decoding is completely completed;
[0036] The output unit is used to output the decoded data.
[0037] As a further improvement, the generating unit further includes a first decoder and a second decoder, wherein the first decoder performs a forward decoding operation and outputs an error code element matrix; the second decoder is used to invert the error rows or columns in turn according to the coordinate information of the error code element matrix, perform decoding operations, and update the error code element matrix.
[0038] Compared with the prior art, the technical solution of the present invention has the following technical effects:
[0039] 1. The present invention only performs iterative error correction on erroneous rows and columns, which can reduce the requirements for hardware resources and improve decoding efficiency;
[0040] 2. By correcting the error of a single row or a single column successfully first, the error matrix converges and approaches the correctable error matrix, and finally all code elements are corrected successfully; for the row or column with 0 uncertain codes, step S3 can be executed once for one iteration to successfully correct the error, and the error matrix converges to the correctable range; for the row or column with 1 uncertain code, step S3 can be executed three times for successful error correction, and the error matrix can be further converged, so that the number of uncertain codes in the row or column is reduced to 0 when step S2 is executed for the second time, and step S3 can be executed once more for the second time to successfully correct the error, and the error matrix converges to the correctable range. Therefore, without increasing the error correction code length, the error correction capability of the row error correction is at least improved to 2X+3, and the error correction capability of the column is improved to 2Y+3. The improved decoding method of this patent can quickly obtain successful decoding within a limited number of iterations, and the hardware complexity is low, and the impact on the decoding performance is small.
[0041] 3. For application scenarios with high channel stability, combined with TPC coding technology, the code length can be shortened, the correctability of the original row / column can be reduced, and then the patented method can be applied to improve the error correction capability. Storage space can be saved without affecting the error correction performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 This is a flowchart of the improved TPC error correction algorithm of the present invention.
[0043] Figure 2 This is a structural block diagram of the improved TPC error correction device of the present invention.
[0044] Figure 3 The hardware structure block diagram of a preferred embodiment of the present invention.
[0045] Figure 4 It is a flowchart of the third decoder in the present invention.
[0046] Figure 5 It is a schematic diagram of the decoding process of the preferred embodiment of the present invention.
[0047] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION
[0048] The technical solution provided by the present invention will be further described below in conjunction with the accompanying drawings.
[0049] See also Figure 1 , which is a flowchart of an improved TPC error correction algorithm of the present invention, at least includes the following steps:
[0050] Step S1: Obtain coded data and execute TPC decoding algorithm to generate an error symbol matrix;
[0051] Step S2: randomly selecting a row or column of data in the error symbol matrix, preferably selecting the row or column with the least number of uncertain symbols as the data to be decoded first;
[0052] Step S3: modifying part of the code elements of the row or column according to the number of uncertain code elements, performing decoding operation on the row or column until decoding is successful, and updating the error code element matrix;
[0053] Step S4: Repeat the above steps until the error symbol matrix has no recorded data, and output the decoded data. In step S4, steps S2 to S3 are repeated until all rows or columns are decoded successfully. Alternatively, in step S4, steps S2 to S3 are repeated until the error symbol matrix reaches the error correction capability of the TPC decoding algorithm, and then step S1 is performed, which can significantly improve the decoding efficiency.
[0054] In the above step S1, on the basis of pre-decoding, the data of successful and failed error correction are separated and marked, and the position information of the cross-code elements of the error rows and columns is extracted to form an error code element matrix. In the subsequent decoding process, only the rows / columns that have not been successfully corrected are deeply corrected, saving decoding time. By correcting the errors of a single row / column first, the error matrix converges and approaches the correctable error matrix, and finally all code elements are successfully corrected; for application scenarios with high channel stability, the present invention combines TPC coding technology to appropriately shorten the code length without affecting the error correction capability, thereby saving storage space.
[0055] The present invention proposes a method strategy for single row / column limit error correction. Assuming that the number of uncertain code elements is K, step S3 further performs the following steps:
[0056] Step S31: Select K+1 symbols in the selected row or column according to the position information in the error matrix;
[0057] Step S32: performing a negation operation on the selected K+1 code elements;
[0058] Step S33: Execute the TPC decoding algorithm on the row or column after the codeword is modified. If the decoding is unsuccessful, perform all inversion operations on the selected row or column according to the position information in the error matrix, and then execute the TPC decoding algorithm on the selected row or column again. If it is still unsuccessful, re-execute step S31 until all iterative processes are completed; if the decoding is successful, execute step S4.
[0059] In the above technical solution, when the uncertain codeword K of a single row or a single column in the error matrix is 0, the step S31 is: invert any one codeword of the selected row or column according to the position information of the error matrix, and then execute step S33 once to successfully correct the error of the row or the column.
[0060] As a further improvement, when the uncertain code element K in a single row or column in the error matrix is 1, the specific steps are as follows:
[0061] When step S31 is executed for the first time, two code elements are randomly selected from the selected row or column and inverted according to the position information of the error matrix. When the two code elements are of the same nature, the error correction can be successfully performed by executing step S33 once; when the two code elements are of different natures, the first execution of step S33 fails and step S31 needs to be repeated.
[0062] When executing step S31 for the second time, select one of the code elements selected for the first time, and then select another code element according to the position of the error matrix to invert them together. When the two code elements selected for the second time are code elements of the same nature, the second step S33 can successfully correct the error; when the two code elements are code elements of different natures, the second execution of step S33 fails. At this time, it can be determined that the other code element selected for the first time and the newly selected code element for the second time are code elements of the same nature. When executing step S31 for the third time, invert these two code elements together, which can ensure that the third step S33 is successfully corrected.
[0063] By adopting the technical solution of the present invention, in order to reduce the number of decoding iterations in step S3, it is necessary to reduce the number of uncertain codewords as much as possible. In a preferred embodiment, in step S1, a positive decoding operation is first performed, and the positive decoding operation is a traditional TPC decoding process. If the decoding is successful, the decoding result is directly output, otherwise the error codeword matrix is output. Take a simple two-dimensional TPC decoding structure as an example, where the number of codewords after each row encoding is M, the number of codewords after each column encoding is N, and its row decoding error correction capability is X codewords, and the column decoding capability is Y codewords. The above decoding order is to perform row decoding first, then column decoding, or to perform column decoding first, then row decoding. For correctable rows and columns, the decoder outputs the corrected information codeword; for uncorrectable rows and columns, the original code is output. After multiple iterations of the positive decoding operation, if the decoding cannot be completed, an error codeword matrix is formed according to the position information of the cross codewords of the error rows and columns, where the number of error rows is b1 and the number of error columns is a1. At this time, the number of uncertain code elements in each row is b1-X, and the number of uncertain code elements in each column is a1-Y.
[0064] Furthermore, step S1 includes a negative decoding operation, which performs the following steps:
[0065] Obtain the error code element matrix output after the positive decoding operation;
[0066] According to the position information in the error matrix, the error rows or columns are inverted in sequence and a decoding operation is performed;
[0067] If the decoding is successful, the decoding result is directly output, otherwise the error symbol matrix is updated and output. The number of error rows is b2, and the number of error columns is a2.
[0068] After multiple iterations of the negative decoding operation, the coding error correction of the columns less than or equal to 2X+1 or the rows less than or equal to 2Y+1 output by the positive decoding can be successfully performed. At this time, the row error correction capability can be increased to 2X+1 code elements, and the column error correction capability can be increased to 2Y+1 code elements. When the error matrix is formed after both the positive decoding and the negative decoding fail, it can be judged that the good codes and bad codes in each row of the error matrix are greater than X, and the number of uncertain code elements is a2-2X-2, and the good codes and bad codes in each column are greater than Y, and the number of uncertain code elements is b2-2Y-2. This reduces the number of uncertain code elements as much as possible. For the case where the number of uncertain codes is relatively small, the decoding strategy of the present invention can reduce the number of iterations and reduce the impact on the decoding performance.
[0069] See also Figure 2 , which is a principle block diagram of an improved TPC error correction device of the present invention, at least comprises:
[0070] A first acquisition unit, used to acquire coded data;
[0071] A generating unit, configured to execute a TPC decoding algorithm on the coded data to generate an error symbol matrix;
[0072] A second acquisition unit is used to acquire data of any selected row or column in the error symbol matrix;
[0073] A decoding unit, used for modifying part of the code elements of the selected row or column according to the number of uncertain code elements, and performing a decoding operation on the row or column until the decoding is successful to update the error code element matrix;
[0074] A judging unit, used to judge whether the decoding is completely completed;
[0075] The output unit is used to output the decoded data.
[0076] In a preferred embodiment, the generation unit further includes a first decoder and a second decoder, wherein the first decoder performs a forward decoding operation and outputs an error symbol matrix; the second decoder is used to perform a decoding operation after inverting the error rows or columns in turn according to the coordinate information of the error symbol matrix, and update the error symbol matrix. In the present invention, each decoder can control the number of sub-iterative decodings through a hardware processor. The decoder in the present invention can be a multiplexed resource or an independent resource. In a preferred embodiment, a first judgment is also performed before the second acquisition unit is executed, which is mainly used to plan the next round of decoding strategies, such as determining which row or column of the error symbol matrix to start from, and determining how to invert the next step.
[0077] The technical solution of the present invention is described in detail below through a specific embodiment. Figure 3, which is a hardware structure diagram of a preferred embodiment of the error correction algorithm of the present invention, including four decoding processes, firstly, the uncorrectable error matrix is located by positive decoding of the first decoder and negative decoding of the second decoder, and then the single row / column codewords in the error matrix are corrected first, so that the error matrix converges to the correctable range, and finally all codewords are corrected successfully. Of course, the decoding structure currently described is only an optional structure of this patent.
[0078] The coded data buffer receives the coded information codewords. Since the transmission bit widths of different signals may be different and may not be consistent with the row or column grouping code widths of the decoder, the coded information is first transmitted to the coded data buffer. The decoders described in this patent all support obtaining information codewords of a specified bit width from the data buffer in a pingpong manner. This patent can be set through a hardware processor to support decoding of different code lengths.
[0079] Since TPC coding is independent between rows and columns, it also supports independent decoding. The first, second, and third decoders in this patent can support multiple decoding engines and execute decoding of each row / column concurrently to improve decoding efficiency; the main decoding processes in this patent can also be in parallel mode. When deep error correction (second decoder and third decoding) is performed in this round of decoding, new decoding can be started without waiting for all decoding to be completed before starting new decoding, thereby improving decoding efficiency.
[0080] The first decoder is a traditional TPC decoding process. Taking a simple two-dimensional TPC decoding structure as an example, the number of code elements after each row encoding is M, the number of code elements after each column encoding is N, and its row decoding error correction capability is X code elements, and the column decoding capability is Y code elements. The decoding order in the description is to perform row decoding first, then column decoding, or to perform column decoding first, then row decoding. For correctable rows and columns, the decoder outputs the corrected information code elements; for uncorrectable rows and columns, the original code is output. If the first decoder successfully corrects the error, the first round of decoding is directly transmitted to the decoding data buffer, and the decoding is marked as successful; if the first decoder still fails to correct the error after multiple sub-iterations, the first round of decoding results are first transmitted to the intermediate data buffer, and the row and column mark bits are added to the data of the erroneous rows and columns and transmitted to the first data buffer.
[0081] The hardware processor can control the number of sub-iterations of the first decoder and record the error correction status of the rows and columns during the sub-iteration process, including the error row number, error column number, number of successfully corrected code elements, and other information, and form an error code element matrix A1. If the number of rows that failed to correct is b1, and the number of columns that failed to correct is a1, then the area of the error matrix is b1*a1. When a1≤X or b1≤Y, the first round of sub-iterative decoding can successfully correct the error; when a1>X and b1>Y, the area of the error matrix constructed at this time is larger than the correctable range supported during encoding, and the first round of decoding error correction fails. The hardware processor starts the second decoder.
[0082] The second decoder obtains the data of a single row / column from the first data buffer, inverts the data according to the code element coordinate information of the error matrix (negative decoding), and determines whether to perform decoding or column decoding according to the row and column mark bit information. Since the information code element is binary, after the code element is inverted, the original error code element will become the correct code element, and the original correct code element will become the error code element. After the new code element enters the next decoding, the error code elements of some rows and columns may be reduced, thereby achieving the correctability of the row / column. The second decoder only decodes the erroneous rows and columns, which can reduce the single iterative decoding from M+N times to a1+b1 times, saving decoding time to a great extent.
[0083] The hardware processor can control the number of sub-iterations of the second decoder and record the error correction status of the rows and columns during the sub-iteration process, including the error row number, error column number, number of successfully corrected symbols, and other information, and form an error symbol matrix A2. The number of rows that failed to correct is recorded as b2, and the number of columns that failed to correct is recorded as a2. At this time, the area of the error matrix is b2*a2, and the judgment is made:
[0084] 1. When a2=0 or b2=0, iterative decoding is no longer performed, the decoding is terminated, the decoding result is output, and the data in the original intermediate data buffer is fed back to the decoding data buffer, and the decoding is marked as successful;
[0085] 2. When 0<a2≤X or 0<b2≤Y, the second round of sub-iterative decoding is performed, and the error correction can be successful after the next sub-iterative decoding;
[0086] 3. When X<a2≤2X+1 or Y<b2≤2Y+1, the codewords in the new error codeword matrix are inverted again, and then the second round of sub-iterative decoding is performed. The error correction can be successful after the next sub-iterative decoding;
[0087] 3. When a2>2X+1 and b2>2Y+1, combined with the previous round of decoding results, the number of good codes and error codes in the error row is both>X, and the number of good codes and error codes in the error column is both>Y. Simple negative decoding operations can no longer break through the error correction capability of TPC. At this time, the hardware processor will mark the second decoder as failed to correct the error, transmit the second round of decoding results to the second data buffer, and notify the start of the third decoder.
[0088] Combining the above two rounds of decoding, the error situations of row error symbol ≤ 2X+1 or column error symbol ≤ 2Y+1 can be successfully corrected, and the correctable error matrix A2 is redefined to meet the conditions of a2≤2X+1 and b2≤2Y+1.
[0089] The third decoder is the core of the present invention, which can only perform error correction on the erroneous rows and columns. The specific process is as follows: Figure 4 . Since the first two decoders failed in error correction, it can be determined that the good codes and bit errors in each row of the error matrix are greater than X, and the number of uncertain code elements is a2-2X-2; the good codes and bit errors in each column are greater than Y, and the number of uncertain code elements in each column is b2-2Y-2. The third decoder needs to perform extreme error correction on a single row / column first. An optional solution is to find a2-2X-1 code elements with the same properties in a specified row or b2-2Y-1 code elements with the same properties in a specified column, invert these code elements, and then perform sub-iterative decoding to achieve error correction for the row.
[0090] When a2=2X+2 or b2=2Y+2, the good codes and error codes in the row or column are equal and in a balanced state, and the number of uncertain code elements in the row or column is 0. At this time, as long as a code element in a row / column in the error matrix is inverted, if the original nature of the code element is an error code element, one positive decoding can successfully correct the error of the row / column; if the original nature of the code element is a correct code element, one decoding cannot successfully correct the error, and it is necessary to invert the code element of the row / column according to the position information of the error matrix and then perform one decoding, and the row / column can be successfully corrected. After the row / column is successfully corrected, a new error matrix A3 is formed, which meets the conditions a3=2X+1, b3=2Y+1, and can reach the error correction capability range of the second decoder.
[0091] Further, when a2=2X+3 or b2=2Y+3, at this time, the number of uncertain code elements in the row and column is 1. In theory, it is only necessary to find 2 code elements of the same nature in a row / column, invert them, and perform step S33 decoding once, and the row / column can be successfully corrected. If the row error correction is performed first, the row can be successfully corrected by performing the sub-iteration row decoding of step S3 at most 3 times, forming a new error matrix information. At this time, the number of uncertain code elements in the error column is reduced to 0. According to the position information of the new error matrix, find any code element in the selected column to invert, and then perform step S33 column decoding operation once, then the column can be successfully corrected. At this time, a new error matrix A3 is formed, which satisfies the conditions a3=2X+1, b3=2Y+1, and can reach the error correction capability range of the second decoder.
[0092] At this time, the third round of decoding results are transmitted to the third data buffer, and the second decoder is notified to perform decoding. The completion condition of the third decoder can also be set to complete error correction, so the second execution of the second decoder can be omitted at this time. The third decoder can preset a threshold of the number of sub-iterations. When the number of error corrections exceeds the threshold, the error correction of the third decoder is actively terminated and the main controller is notified.
[0093] To summarize, the third decoder only needs 1 round of iterative decoding to increase the row error correction capability to 2X+2, and the column error correction capability to 2Y+2; if at most 4 rounds of iterative decoding are added, the row error correction capability can be increased to 2X+3, and the column error correction capability can be increased to 2Y+3.
[0094] Decoding example Figure 5, assuming that the original error correction capability of the TPC decoder is 2 codewords per row and 3 codewords per column. The uncorrectable error matrix A(9,6) is constructed by one positive decoding and one negative decoding, where the number of rows with failed error correction is 9 and the number of columns with failed errors is 6. It can be inferred that the number of correct codewords in each row of the error matrix is 3, the number of error codewords is 3, and the number of uncertain codewords is 0; the number of correct codewords in each column is greater than or equal to 4, the number of error codewords is greater than or equal to 4, and the number of uncertain codewords is 1. According to the number of uncertain codewords in rows and columns, single-row iterative error correction is preferred. Invert the first codeword in the first row of the error matrix A, and then perform decoding of the row. If the error correction is successful, it can be inferred that the codeword was originally an error codeword. After inversion, the error correction is successful. At this time, the number of error codewords in the row is 2, which meets the original row decoding capability of TPC. If the error correction fails, it can be inferred that the codeword was initially a correct codeword, and became an error codeword after being negated. At this time, the number of error codewords in this row is 4, which does not meet the original row decoding capability of TPC. Then if all the 6 codewords marked in the error matrix of this row are negated, 4 correct codewords and 2 error codewords can be obtained. At this time, decoding is performed again, and the error correction of this row can be successful; at this time, a new error matrix B(8,6) is formed, and the number of uncertain codewords in the row and column is 0. You can choose to perform error correction on a single row or a single column again, using the same method as above. In this example, the second row is successfully corrected. At this time, a new error matrix C(7,6) is formed. At this time, one of the correct codewords or error codewords in each column must be less than 4, so successful error correction can be achieved by performing positive decoding or negative decoding on a single column. At this point, all the original error codewords have been successfully corrected.
[0095] The above embodiments are only used to help understand the method and core idea of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.
[0096] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. An improved TPC error correction algorithm, characterized in that: At least the following steps are included: Step S1: Obtain coded data and execute TPC decoding algorithm to generate an error symbol matrix; Step S2: randomly selecting a row or column of data in the error symbol matrix; Step S3: modify some code elements of the row or column, perform decoding operation on the row or column until decoding is successful, and update the error code element matrix; Step S4: Repeat the above steps until there is no erroneous data in the error symbol matrix, and output the decoded data; In step S3, the code elements in the corresponding row or column are modified according to the number of uncertain code elements in the row or column; assuming that the number of uncertain code elements is K, the following steps are performed: Step S31: According to the position information of the error matrix, K+1 symbols are randomly selected in the selected row or column; Step S32: performing a negation operation on the selected K+1 code elements; Step S33: Execute the TPC decoding algorithm on the row or column after the modified codeword. If the decoding is unsuccessful, perform all inversion operations on the selected row or column according to the position information of the error matrix and then execute the TPC decoding algorithm again. If the decoding is still unsuccessful, re-execute step S31 until all iterative processes are completed. If the decoding is successful, execute step S4; When the uncertain symbol K of a single row or column in the error matrix is 0, the step S31 is: According to the position information of the error matrix, one code element of the selected row or column is inverted; Execute step S33 again to successfully correct the error of the row or column; When the uncertain code element K of a single row or column in the error matrix is 1, the specific steps are as follows: When step S31 is executed for the first time, two code elements are randomly selected from the selected row or column according to the position information of the error matrix to be inverted. When the two code elements are code elements of the same nature, step S33 is executed to successfully correct the error. When the two code elements are code elements of different natures, the first execution of step S33 fails, and step S31 needs to be executed again; When executing step S31 for the second time, select one of the code elements selected for the first time, and then select another code element according to the position of the error matrix to invert them together. When the two code elements selected for the second time are code elements of the same nature, execute step S33 again to successfully correct the error; when the two code elements are code elements of different natures, the second execution of step S33 fails, thereby judging that the other code element selected for the first time and the newly selected code element for the second time are code elements of the same nature. When executing step S31 for the third time, invert these two code elements at the same time, and the third step S33 will successfully correct the error.
2. The improved TPC error correction algorithm according to claim 1, characterized in that: Step S1 includes performing a positive decoding operation and a negative decoding operation. The positive decoding operation is a traditional TPC decoding process. If the decoding is successful, the decoding result is directly output, otherwise the original codeword and the error codeword matrix are output; The negative decoding operation performs the following steps: Obtain the error code element matrix output after the positive decoding operation; According to the position information of the error matrix, the error rows or columns are inverted in sequence and a decoding operation is performed; If a row or column is decoded successfully, the decoding result is output, and the error symbol matrix is updated and output at the same time; Otherwise, the original codeword and error codeword matrix are output.
3. The improved TPC error correction algorithm according to claim 1, characterized in that: In step S2, the row or column with the least number of uncertain code elements is selected for error correction first.
4. The improved TPC error correction algorithm according to claim 1, characterized in that: In step S4, steps S2 to S3 are repeatedly executed until all rows or columns are decoded successfully.
5. The improved TPC error correction algorithm according to claim 1, characterized in that: In step S4, steps S2 to S3 are repeatedly executed to make the error symbol matrix within the error correction capability range of the TPC decoding algorithm, and then step S1 is executed.
6. An improved TPC error correction device according to any one of the algorithms of claims 1 to 5, characterized in that: At least: A first acquisition unit, used to acquire coded data; A generating unit, configured to execute a TPC decoding algorithm on the coded data to generate an error symbol matrix; A second acquisition unit is used to acquire data of any selected row or column in the error symbol matrix; A decoding unit, used to modify part of the code elements of the selected row or column, and perform a decoding operation on the row or column until the decoding is successful to update the error code element matrix; A judging unit, used to judge whether the decoding is completely completed; The output unit is used to output the decoded data.
7. The improved TPC error correction device according to claim 6, characterized in that: The generating unit further comprises a first decoder and a second decoder, wherein the first decoder performs a forward decoding operation and outputs an error symbol matrix; the second decoder is used to perform a decoding operation after successively inverting the error rows or columns according to the position information of the error matrix, and update the error symbol matrix.
