JP2010034887A - Encoding apparatus, encoding method and encoding program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an encoding apparatus capable of reducing computational complexity. <P>SOLUTION: The encoding apparatus for linear codes includes: a parity generating section 11 for generating at least one parity from first data as input data by using a generation matrix and for output of first encoded data; a row selecting section 13 for selecting a row, in which the number of parity which is not generated out of parities of the input second data corresponding to non-zero elements of rows is one, from all rows of a check matrix of linear codes; a disappearance correcting section 14 for generating a parity corresponding to the relevant row by performing disappearance correction processing and for output of second encoded data; and a first selecting section 15 for output of either the first encoded data or the second encoded data as second data, wherein until linear encoding is finished, the disappearance correcting section 14 outputs the output second encoded data to the first selecting section 15, and when linear encoding is completed, the disappearance correcting section outputs the output second encoded data as a code word, thereby generating the code word. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an encoding method and an encoding device for a linear code such as an LDPC (Low Density Parity Check) code.

  In recent years, LDPC codes have been considered promising as codes having excellent correction capability (for example, Non-Patent Documents 1 to 3).

  When video or audio streaming transmission is performed on an IP network, a lost packet may be sent by UDP (User Datagram Protocol) in order to maintain real-time performance. In such a case, if there is a section for wireless transmission on the transmission path, it is necessary to take measures against lost packets. As a countermeasure against this, it is conceivable to interleave the packets and perform error correction coding. As an error correction code to be used, an LDPC code that easily increases the interleave length is also a promising candidate.

  For example, as shown in FIG. 5, p × k bits of data are made into blocks of p bits, k bits of data obtained from each block are encoded, and (n−k) bits of parity are added. As a result, a total of p × (n−k) bits of parity is added, and these parities form an (n−k) block. At the time of transmission, header data (such as a UDP header / IP header) is added to each block as data of one packet to form an IP packet. Interleaving may be performed by collecting a plurality of bits instead of one bit from one block and encoding them into one codeword, or collecting and encoding one bit from a plurality of blocks.

  Here, the error correction code to be used is described as a systematic code that only adds parity without changing data. Compared to a non-systematic code in which data is converted, the systematic code has an advantage that a received packet can be used as it is even when a lost packet cannot be corrected and restored.

  Even if n packets are not prepared after transmission, if the lost packet data is lost and error correction processing is performed on all of the p code words to restore one bit at a time, the original block Can be restored. IP packets have a maximum packet length of 1,500 bytes in Ethernet (registered trademark), the IP header length is 20 bytes as standard, and the UDP header length is 8 bytes, so the data is a maximum of 1,472 bytes. = 11,776 bits can be used. When streaming transmission is performed, it is required to increase the amount of data included in one packet as much as possible and to reduce the transmission rate increased by the header as much as possible and efficiently transmit the data. It is considered to be about 10,000 bits. When one IP packet is converted into one symbol of the code, the addition of the symbols performs an exclusive OR of about 10,000 bits.

A linear code can define a code structure by a generator matrix G of k rows and n columns and a check matrix H of (n−k) rows and n columns where G · H ^T = 0. And H · G ^T = 0 may be described. The encoding process is performed by multiplying the data (vector) D = (d ₀ , d ₁ ,..., D _(k−1) ) by the generator matrix G, and the code word (vector) C = (c ₀ , c ₁ ,..., C _(n-1) ) = D · G. Here, when C · H ^T is calculated for the code word, C · H ^T = D · G · H ^T = D · 0 = 0.

In addition, the Reed-Solomon code, which is the maximum distance separation code and is often used in a recording apparatus or the like, can be encoded by performing a correction process with an uncalculated parity as erasure.
RG Gallager, “Low density parity check codes”, in Research Monograph series. Cambridge, MIT Press (1963) DJ MacKay, “Good error-correcting codes based on very sparse matrices”, IEEE Trans. Inform. Theory, vol.45, pp.399-431 (1999) RM Tanner, “A recursive approach to low complexity codes”, IEEE Trans. Inform. Theory, vol.27, pp.533-547 (1981)

  However, the conventional encoding process has a problem that the calculation amount is enormous.

  That is, in the actual encoding process, when the data D is to be encoded by applying the generator matrix G, even if the codeword symbol is an element of the Galois field GF (2), the calculation amount is approximately ( It is necessary to add symbols n−k) × (k / 2-1) times. Addition of one symbol is an exclusive-or process of about 10,000 bits, and the amount of calculation is enormous. Therefore, encoding processing time and processing resources are required. In addition, a normal LDPC code is not a maximum distance separation code, and a method of encoding by performing correction processing with an uncalculated parity as an erasure cannot be used.

  In order to solve the above problems, an object of the present invention is to provide an encoding device capable of reducing the amount of calculation.

  In order to achieve the above object, an encoding apparatus according to the present invention is an encoding apparatus that generates a code word by linear encoding for input data, and the input data is input as first data, Parity generation for generating at least one parity from the first data using the generation matrix of the linear code, and outputting first encoded data in which the generated parity is added to the first data And the second data is input, and among the parity of the second data corresponding to the non-zero element of the row out of all the rows of the parity check matrix of the linear code, A row selection unit that selects one row and a row selected by the row selection unit are used to generate the parity that has not been generated by performing a correction process as erasure, and the generated parity is And an erasure correction unit that outputs the second encoded data in addition to the data, and when the first encoded data is input, the first encoded data is output as the second data. A first selection unit that outputs the second encoded data as the second data when the second encoded data is input, and the erasure correction unit performs linear encoding The output second encoded data is output to the first selection unit until the end, and when the linear encoding is completed, the output second encoded data is output as the code word. Thus, the code word is generated.

  According to this, the erasure correction unit can generate parity by erasure correction processing with a small calculation amount, and the calculation amount can be reduced.

  Further, when the second data is input, the second selection unit that outputs the second data as the first data to the parity generation unit, and the row selection unit select the second data. When it is determined that there is no row to be selected by the row selection unit and a row that the row selection unit selects, the first selection unit determines whether or not there is a row of the parity check matrix. A switching unit that switches the second selection unit to output the second data, and the parity generation unit receives the second selection when the second data is input to the second selection unit. At least one parity is generated from the first data input from the unit, and first encoded data in which the generated parity is added to the first data is output.

  According to this, when the erasure correction unit cannot perform the erasure correction process, the parity generation unit generates a parity. As a result, the erasure correction unit can perform erasure correction processing, and the amount of calculation can be reduced.

  Further, the parity generated by the parity generation unit may be determined so that the row selection unit selects at least one row of the parity check matrix.

  According to this, the parity generated by the parity generation unit is determined in advance so that the row selection unit can select a row. As a result, the erasure correction unit can perform erasure correction processing, and the amount of calculation can be reduced.

  In addition, the present invention can be realized not only as such an encoding apparatus, but also as a method using a processing unit constituting the apparatus as a step. Furthermore, the present invention can be realized as a program for causing a computer to execute these steps, or can be realized as a recording medium such as a computer-readable CD-ROM on which the program is recorded, or as information, data, or a signal indicating the program. It can also be realized. These programs, information, data, and signals may be distributed via a communication network such as the Internet.

  According to the present invention, parity can be generated by erasure correction processing with a small calculation amount, and the calculation amount of the entire encoding process can be greatly reduced.

  Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment 1)
FIG. 1 is a block diagram showing an example of a functional configuration of an encoding device 10 according to Embodiment 1 of the present invention.

  As shown in FIG. 1, the encoding apparatus 10 of Embodiment 1 includes a parity generation unit 11, a check matrix determination unit 12, a row selection unit 13, an erasure correction unit 14, a first selection unit 15, and a second selection. A unit 16 and a switching unit 17 are provided. The encoding device 10 is a computer that includes a CPU (Central Processing Unit) and the like, inputs necessary data into a memory, executes a program, and outputs data.

  The parity generation unit 11 receives the first data, generates at least one parity from the first data by using a linear code generation matrix, and the generated parity is added to the first data. The first encoded data is output. In addition, when the second data is input to the second selection unit 16, the parity generation unit 11 generates at least one parity from the first data input from the second selection unit 16, The first encoded data obtained by adding the generated parity to one data is output.

  The row selection unit 13 receives the second data, and among the parity of the second data corresponding to the non-zero element of the row, the parity that has not been generated is selected from all the rows of the linear code parity check matrix. Select one row.

  The parity check matrix determination unit 12 determines whether there is a row of the parity check matrix that can be selected by the row selection unit 13.

  The erasure correction unit 14 uses the row selected by the row selection unit 13 to generate the parity that has not been generated as one erasure and performs correction processing, and adds the generated parity to the second data. To output second encoded data. Specifically, the erasure correction unit 14 is not generated because the value obtained by multiplying the row composed of the second data and the parity by the transpose matrix of the row selected by the row selection unit 13 becomes zero. Parity is calculated and generated.

  The first selection unit 15 outputs the first encoded data as the second data when the first encoded data is input, and the first selector 15 when the second encoded data is input. 2 encoded data is output as second data.

  The second selection unit 16 outputs the data to the parity generation unit 11 as the first data when data is input to the encoding device 10, and the second when the second data is input. Are output to the parity generator 11 as first data.

  When the check matrix determination unit 12 determines that the row selected by the row selection unit 13 does not exist, the switching unit 17 outputs the second data to the second selection unit 16 by the first selection unit 15. Switch as follows.

  Here, each data such as the first data, the second data, the first encoded data, and the second encoded data is output from each processing unit and then written to the memory. The necessary data is read from the memory and each process is performed.

  FIG. 2 is a flowchart showing the encoding process executed by the encoding apparatus 10.

  First, data for encoding is input to the encoding device 10 (S102).

  Then, the second selection unit 16 selects the data input to the encoding device 10 as first data, and outputs the first data to the parity generation unit 11 (S104).

  The parity generation unit 11 generates at least one parity from the input first data by using a linear code generation matrix, and the first encoding in which the generated parity is added to the first data Data is output (S106).

  Next, the first selection unit 15 selects the first encoded data output from the parity generation unit 11 as second data, and outputs the second encoded data to the parity check matrix determination unit 12 (S108).

  The parity check matrix determination unit 12 determines whether there is a row of the parity check matrix that can be selected by the row selection unit 13 (S110).

  When the check matrix determination unit 12 determines that there is no row in the check matrix (NO in S110), the switching unit 17 causes the first selection unit 15 to output the second data to the second selection unit 16. Switch as follows. Thereby, the second selection unit 16 selects the second data as the first data, and outputs it to the parity generation unit 11 (S104).

  The parity generation unit 11 generates at least one parity from the first data input from the second selection unit 16, and the first code obtained by adding the generated parity to the first data The converted data is output (S106).

  Then, the first selection unit 15 outputs the first encoded data as second data to the parity check matrix determination unit 12 (S108), and the parity check matrix determination unit 12 determines whether there is a row of the parity check matrix. Is determined (S110).

  Also, when the parity check matrix determination unit 12 determines that there is a row of the parity check matrix (YES in S110), the second data is output to the row selection unit 13, and the row selection unit 13 selects the row of the parity check matrix. (S112). Specifically, the row selection unit 13 selects, from all the rows of the parity check matrix, a row that has one ungenerated parity among the parity of the second data corresponding to the non-zero element of the row. select.

  Then, the erasure correction unit 14 uses the row selected by the row selection unit 13 to generate the parity that has not been generated as one erasure and performs correction processing, and the generated parity is used as the second data. In addition, the second encoded data is output (S114).

  Further, the erasure correction unit 14 determines whether or not the linear encoding of the data input to the encoding device 10 has been completed, that is, whether or not a code word has been generated (S116).

  If the erasure correction unit 14 determines that the linear encoding has not been completed and the code word has not been generated (NO in S116), the erasure correction unit 14 outputs the second encoded data to the first selection unit 15, The first selection unit 15 outputs the second encoded data as the second data (S108).

  Moreover, the erasure | correction correction | amendment part 14 produces | generates a code word by outputting the output 2nd encoded data as a code word, when linear encoding is complete | finished. Thereby, the erasure correction unit 14 determines that a code word has been generated (YES in S116), and the encoding process ends.

  As described above, the erasure correction unit 14 outputs the output second encoded data to the first selection unit 15 until the linear encoding is completed, and is output when the linear encoding is completed. A codeword is generated by outputting the second encoded data as a codeword.

  3A and 3B are diagrams for specifically explaining an example of the encoding process of the encoding device 10.

The encoding apparatus 10 generates parity of r symbols from the input data D = (d ₀ , d ₁ ,..., D _(k−1) ) of _k symbols, and outputs n symbols. Code word C = (d ₀ , d ₁ , d ₂ , d ₃ ,..., D _(k−2) , d _(k−1) , p ₀ , p ₁ ,..., P _(r−2) , p _(r-1) ) is output. Note that the number r of parity symbols is (n−k).

The code generation matrix G is a matrix of k rows and n columns as shown in FIG. 3A, and the first to kth columns from the left (hereinafter referred to as the data position portion) that generate the data portion of the codeword are An element in the u-row and v-column in the (k + 1) -th column to the n-th column from the left (hereinafter referred to as the parity position portion) that is a unit matrix and generates the parity part of the codeword is g _{(u-1 ), (vk-1)} . Also, the check matrix H corresponding to this is a matrix of (n−k) rows and n columns as shown in FIG. 3B. Here, each element of the data D, the generation matrix G, and the check matrix H is a binary value, and an exclusive-or process is performed in the calculation.

The parity generation unit 11 selects q columns from the parity position portion of the generation matrix G as shown in FIG. 3A, and j = 0, 1,..., (Q−1), and the parity p _j = g _{0, j.} d ₀ + g _{1, j} · d ₁ + g _{2, j} · d ₂ + …… + g _{(k-2), j} · d _(k-2) + g _{(k-1), j} · d _(k-1) Generate and output as first encoded data. Here, the selection of q columns from the parity position portion of the generator matrix G is selected such that there are many rows of the check matrix H that can be selected by the row selection unit 13. The selection of the q column may be an arbitrary column.

  In FIG. 3B, in the check matrix H, columns corresponding to the data position portion of the generation matrix G and the generated parity portion are indicated by hatching.

  The parity check matrix determination unit 12 first has one non-zero element in the row of the partial matrix excluding the data position portion and the column corresponding to the already generated parity from the parity check matrix. Determine if a row exists.

  If there is even one row where the number of elements whose values are non-zero is one, the switching unit 17 is switched to the row selection unit 13 side, and the row selection unit 13 changes the data position portion from the parity check matrix. Of the rows of the sub-matrix excluding the columns corresponding to the already generated parity, the row having one nonzero element is selected.

において、まだ生成されていないパリティｐ_q〜ｐ_(r-1)の係数となるｈ_i,(k+q)〜ｈ_i,(n-1)のうち、ｈ_i,jのみが非零である。 If the selected row is the (i + 1) th row and the non-zero element is h _{i, j} (where j is any one of (k + q) to (n−1)), C · H ^T = Expression that extracts only the (i + 1) th line from 0

In, h _i as the coefficient of not yet generated parity _{p q ~p (r-1)} , (k + q) ~h i, (n-1) of, h _{i, j} only in nonzero is there.

により計算し、第２の符号化データとして、第１の選択部１５を経由して、検査行列判断部１２と消失訂正部１４にフィードバックする。 The erasure correction unit 14 _calculates the parity p _(jk) based on the information indicating the row selected by the row selection unit 13.

And is fed back to the parity check matrix determination unit 12 and the erasure correction unit 14 via the first selection unit 15 as the second encoded data.

  Also, in the parity check matrix determination unit 12, when there is no row having one non-zero value element, the switching unit 17 is switched to the parity generation unit 11 side, and the parity generation unit 11, the column corresponding to the parity that has not been generated yet is selected from the parity position portion of the generator matrix G, the parity of the column is generated, and is output again as the first encoded data, and the first selector 15 Is fed back to the check matrix determination unit 12 and the erasure correction unit 14.

  As described above, one parity can be generated by either the erasure correction unit 14 or the parity generation unit 11, and the partial matrix obtained by removing the data position portion and the column corresponding to the already generated parity from the parity check matrix Excludes the column corresponding to the generated parity.

  In this way, the parity check method is selected by the check matrix determination unit 12 and the row selection unit 13, and either the erasure correction unit 14 or the parity generation unit 11 repeatedly calculates the parity, and the remaining parity is calculated. Generate all.

  In an ordinary regular LDPC code, non-zero elements are expected to be almost half in the parity position portion of the generator matrix. Therefore, when generating parity based on the columns of the generator matrix, k / 2) -1 addition in symbol units occurs. On the other hand, since each row of the parity check matrix has (sign order) / (1−k / n) non-zero elements, when parity is generated based on the row of the parity check matrix, one parity is generated. Therefore, ((order of code) / (1-k / n))-2 additions in symbol units occur. Assuming that the value of k is several hundred to several thousand, the range of k / n is about 1/2 to 1/10, and the order of the code is about 3 to 7, parity is generated based on the columns of the generator matrix. In this case, the number of additions per symbol is on the order of hundreds to thousands, and the number of additions per symbol when parities are generated based on the rows of the check matrix is on the order of several tens, which is calculated based on the generation matrix. By calculating a part of the parity based on the parity check matrix, the amount of calculation can be greatly reduced.

  Also, in each row of the parity check matrix, if there is only one element of unknown value among the elements of the received word corresponding to the non-zero element, calculating from the other elements usually corrects the erasure error It is a technique when doing. When correcting a erasure error occurring at an arbitrary position, it is possible to correct a erasure error of about 50 to 90% of the number of parity symbols, although it varies greatly depending on the code. For example, in the case of a code that can correct 90% of erasure errors of the number of parity symbols, if the parity generation unit 11 generates parity of the number of symbols corresponding to 10% that cannot be corrected, the remaining 90% Most (or all) of the parity can be generated by the erasure correction unit 14.

  In this way, the erasure correction unit 14 can generate parity by erasure correction processing with a small amount of calculation. Further, when the erasure correction unit 14 cannot perform the erasure correction process, the parity generation unit 11 generates a parity. As a result, the erasure correction unit 14 can perform the erasure correction process, and the amount of calculation can be reduced.

  If there is even one row in which the number of elements whose values are non-zero is one in the check matrix determination unit 12, the information is output to the row selection unit 13, and the row selection unit 13 May be used as it is (the row selection unit 13 does not substantially exist).

(Embodiment 2)
Next, a second embodiment of the present invention will be described. In Embodiment 1 described above, if there is no row of the parity check matrix H that can be selected by the row selection unit 13, the parity generation unit 11 generates parity. However, in the second embodiment, the row selection unit 13 is configured to select at least one row of the check matrix H.

  FIG. 4 is a block diagram showing an example of a functional configuration of the encoding apparatus according to Embodiment 2 of the present invention.

  As shown in FIG. 4, the coding apparatus 10 of Embodiment 2 includes a parity generation unit 21, a parity generation unit 22, a row selection unit 23, an erasure correction unit 24, and a first selection unit 25 based on a parity check matrix. It has. Similar to encoding apparatus 10 of the first embodiment, encoding apparatus 10 of the second embodiment generates parity of r = (n−k) symbols from input data D of k symbols. Then, a code word of n symbols is output. The generator matrix and the check matrix are also as shown in FIGS. 3A and 3B, as in the first embodiment.

The parity generation unit 21 selects q columns from the parity position portion of the generation matrix G as shown in FIG. 3A, and uses j = 0, 1,..., (Q−1) as the parity p _j = g _{0, j.} d ₀ + g _{1, j} · d ₁ + g _{2, j} · d ₂ + …… + g _{(k-2), j} · d _(k-2) + g _{(k-1), j} · d _(k-1) Generate and output as first encoded data. In FIG. 3B, in the check matrix H, columns corresponding to the data position portion of the generation matrix G and the generated parity portion are indicated by hatching.

The parity generation unit 22 based on the parity check matrix generates the remaining ungenerated parity one symbol at a time. First, in the row selection unit 23, the value of the row of the submatrix H _(nkq) of (n−q−q) columns obtained by removing the data position portion from the parity check matrix and the column corresponding to the already generated parity is the value. Select a row with one non-zero element.

において、まだ生成されていないパリティｐ_q〜ｐ_(r-1)の係数となるｈ_i,(k+q)〜ｈ_i,(n-1)のうち、ｈ_i,jのみが非零である。 If the selected row is the (i + 1) th row and the non-zero element is h _{i, j} (where j is any one of (k + q) to (n−1)), C · H ^T = Expression that extracts only the (i + 1) th line from 0

In, h _i as the coefficient of not yet generated parity _{p q ~p (r-1)} , (k + q) ~h i, (n-1) of, h _{i, j} only in nonzero is there.

により、計算する。 The erasure correction unit 24 _calculates the parity p _(jk) based on the information indicating the row selected by the row selection unit 23.

To calculate.

As described above, since one parity has been generated, it is fed back to the row selection unit 23 and the erasure correction unit 24 as second encoded data via the first selection unit 25. Further, the submatrix excluding the data position portion and the column corresponding to the already generated parity from the parity check matrix is the (n−k−q−1) column portion excluding the column corresponding to the generated parity. The matrix is H _(nkq-1) .

  In this way, the row selection unit 23 selects a row having one nonzero element from the submatrix, and the erasure correction unit 24 selects a row based on the value of the row of the check matrix. By repeatedly calculating the parity corresponding to the zero element, all the parities not generated by the parity generation unit 21 are generated.

Here, the parity generated by the parity generation unit 21 is determined so that the row selection unit 23 selects at least one row of the check matrix H. That is, with respect to q parities generated by the parity generation unit 21, a partial matrix H _x (where x = (n−k−q), (n− In each of k−q−1),..., 2, 1), it is assumed that the number of elements having a non-zero value is preselected so that there is at least one row. .

  With the above configuration, as in the first embodiment, it is possible to replace a part of the parity symbol creation algorithms with erasure correction processing, and for each parity symbol generation, the number of additions per symbol is approximately The number of data symbols can be reduced from (2) -1 times to ((order of code) / (1-coding rate))-2 times, and the calculation amount of the entire encoding process can be greatly reduced. It became possible.

  In this manner, the parity generated by the parity generation unit 21 is determined in advance so that the row selection unit 23 can select a row. As a result, the erasure correction unit 24 can perform the erasure correction process, and the amount of calculation can be reduced.

  The following table shows an example of how many times the addition is actually performed in each of the conventional coding apparatus and the coding apparatus of the present invention. For the number of parity q generated by the parity generation unit 21, a value that minimizes the conventional ratio is selected. From this result, it is clear that according to the present invention, the number of additions can be reduced and the encoding processing speed can be increased.

  Each code has a coding rate (data length / code length) of 3/4 and a parity check matrix generated with random numbers. Therefore, even if the parameters are the same (code length, coding rate, order), the values of q and the number of additions are different if the codes are different, but the tendency is almost the same.

  Note that the parity generation unit 21 selects a row in which the number of elements having a non-zero value in the submatrix is one. However, when the parity generated by the parity generation unit 21 is selected in advance, Since it is also known that the rows are selected in the proper order, the rows may be designated in the already known order without checking the number of non-zero elements.

  In FIG. 3A, the q columns are grouped on the left side of the parity position portion of the parity check matrix, but in reality, when the parity generated by the parity generation unit 21 is selected in advance, it is between r columns of the parity position portion. As the check matrix and generator matrix, they can be used as they are, or the columns can be switched so that they are collected on the left side.

  Further, in the present invention, calculation is described as addition, but this means calculation according to the element of the sign. For example, if the sign is composed of elements of Galois field GF (q), addition of a and b is (a + b) mod q, and if it is composed of elements of Galois field GF (2) Is an exclusive-or. In addition, when a block composed of elements of t-bit Galois field GF (2) is used as one symbol, it means an operation that executes t exclusive-or in parallel. .

  In addition, since the generator matrix G and the check matrix H are unique to each code, if the function of selecting or inputting the information of the generator matrix G and the check matrix H according to the used code is provided, It is obvious that the encoding apparatus 10 that handles a plurality of codes can be configured.

  As described above, the encoding apparatus according to the present invention has been described using the above embodiment, but the present invention is not limited to this.

  That is, the embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  For example, in the present embodiment, each element of the data D, the generator matrix G, and the check matrix H is a binary number, but each element of the data D, the generator matrix G, and the check matrix H is limited to a binary number. Alternatively, it may be a ternary number.

  Further, in the present embodiment, each component such as the parity generation unit 11 is configured by software, but may be configured by dedicated hardware. That is, each functional block in the block diagrams (FIG. 1 and FIG. 4 and the like) may be realized as an LSI that is typically an integrated circuit. Note that these may be individually made into one chip, or may be made into one chip so as to include a part or all of them.

  The encoding apparatus, encoding method, and encoding program according to the present invention can increase the processing speed and are useful for erasure correction such as packet loss during network transmission.

It is a block diagram which shows an example of a functional structure of the encoding apparatus of Embodiment 1 of this invention. It is a flowchart which shows the process of the encoding which an encoding apparatus performs. It is a figure explaining an example of an encoding process of an encoding apparatus concretely. It is a figure explaining an example of an encoding process of an encoding apparatus concretely. It is a block diagram which shows an example of a functional structure of the encoding apparatus of Embodiment 2 of this invention. It is explanatory drawing of transmission packet generation.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Encoding device 11, 21 Parity generation part 12 Check matrix judgment part 13, 23 Row selection part 14, 24 Erasure correction part 15, 25 First selection part 16 Second selection part 17 Switching part 22 Based on check matrix Parity generator

Claims (5)

An encoding device that generates a code word by linear encoding for input data,
The input data is input as first data, at least one parity is generated from the first data using the generation matrix of the linear code, and the generated parity is added to the first data. A parity generation unit that outputs the first encoded data;
Second data is input, and one of the parity of the second data corresponding to the non-zero element of the row is not generated from all the rows of the parity check matrix of the linear code. A line selection section for selecting a line;
Using the row selected by the row selection unit, the parity that has not been generated is corrected and generated by erasure, and the generated parity is added to the second data to generate second encoded data. Erasure correction unit that outputs
When the first encoded data is input, the first encoded data is output as the second data, and when the second encoded data is input, the second code is output. And a first selection unit that outputs the digitized data as the second data,
The erasure correction unit outputs the output second encoded data to the first selection unit until linear encoding is completed, and outputs the second encoded data when linear encoding is completed. The encoding device generates the code word by outputting the encoded data as the code word. further,
A second selector that outputs the second data to the parity generator as the first data when the second data is input;
A parity check matrix determination unit that determines whether or not there is a row of the parity check matrix that can be selected by the row selection unit;
A switching unit that switches the first selection unit to output the second data to the second selection unit when it is determined that there is no row to be selected by the row selection unit;
The parity generation unit generates at least one parity from the first data input from the second selection unit when the second data is input to the second selection unit; The encoding apparatus according to claim 1, wherein the first encoded data in which the generated parity is added to the first data is output. The coding apparatus according to claim 1, wherein the parity generated by the parity generation unit is determined such that the row selection unit selects at least one row of the parity check matrix. An encoding method for generating a code word by linear encoding for input data,
A computer inputs the input data as first data, generates at least one parity from the first data using a generation matrix of the linear code, and generates the generated parity for the first data. A parity generation step of outputting the first encoded data to which
The computer selects a row in which all of the parity of the input second data corresponding to the non-zero element of the row is one not generated among all the rows of the check matrix of the linear code. A row selection step to select;
The computer uses the row selected in the row selection step to generate the parity that has not been generated by performing correction processing as an erasure, and adds the generated parity to the second data to generate the second data. Erasure correction step for outputting encoded data of
The computer outputs the first encoded data as the second data when the first encoded data is input, and the first encoded data when the second encoded data is input. A first selection step of outputting two encoded data as the second data,
In the erasure correction step, the output second encoded data is output to the first selection unit until linear encoding is completed, and the second encoded data output when linear encoding is completed. The encoding method is characterized in that the code word is generated by outputting the encoded data as the code word. A program for generating codewords by linear encoding for input data,
The input data is input as first data, at least one parity is generated from the first data using the generation matrix of the linear code, and the generated parity is added to the first data. A parity generation step of outputting the first encoded data;
Second data is input, and one of the parity of the second data corresponding to the non-zero element of the row is not generated from all the rows of the parity check matrix of the linear code. A row selection step for selecting a row;
Using the row selected in the row selection step, the non-generated parity is generated by performing correction processing as an erasure, and the generated parity is added to the second data to perform the second encoding. Erasure correction step for outputting data;
When the first encoded data is input, the first encoded data is output as the second data, and when the second encoded data is input, the second code is output. Causing the computer to execute a first selection step of outputting the digitized data as the second data,
In the erasure correction step, the output second encoded data is output to the first selection unit until linear encoding is completed, and the second encoded data output when linear encoding is completed. The codeword is generated by outputting the encoded data of the codeword as the codeword.

Images