+

WO2008016117A1 - Procédé de génération de matrice d'inspection, procédé d'encodage, dispositif et système de communication et encodeur - Google Patents

Procédé de génération de matrice d'inspection, procédé d'encodage, dispositif et système de communication et encodeur Download PDF

Info

Publication number
WO2008016117A1
WO2008016117A1 PCT/JP2007/065195 JP2007065195W WO2008016117A1 WO 2008016117 A1 WO2008016117 A1 WO 2008016117A1 JP 2007065195 W JP2007065195 W JP 2007065195W WO 2008016117 A1 WO2008016117 A1 WO 2008016117A1
Authority
WO
WIPO (PCT)
Prior art keywords
matrix
column
parity check
mask
check matrix
Prior art date
Application number
PCT/JP2007/065195
Other languages
English (en)
French (fr)
Inventor
Wataru Matsumoto
Rui Sakai
Hideo Yoshida
Original Assignee
Mitsubishi Electric Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Mitsubishi Electric Corporation filed Critical Mitsubishi Electric Corporation
Priority to JP2008527792A priority Critical patent/JP4901871B2/ja
Priority to EP07791869A priority patent/EP2053751A4/en
Priority to US12/376,344 priority patent/US20100058140A1/en
Priority to CA002660060A priority patent/CA2660060A1/en
Publication of WO2008016117A1 publication Critical patent/WO2008016117A1/ja

Links

Classifications

    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03MCODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/03Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
    • H03M13/05Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits
    • H03M13/11Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits using multiple parity bits
    • H03M13/1102Codes on graphs and decoding on graphs, e.g. low-density parity check [LDPC] codes
    • H03M13/1148Structural properties of the code parity-check or generator matrix
    • H03M13/1177Regular LDPC codes with parity-check matrices wherein all rows and columns have the same row weight and column weight, respectively
    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03MCODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/03Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
    • H03M13/05Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits
    • H03M13/13Linear codes
    • H03M13/19Single error correction without using particular properties of the cyclic codes, e.g. Hamming codes, extended or generalised Hamming codes
    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03MCODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/65Purpose and implementation aspects
    • H03M13/6508Flexibility, adaptability, parametrability and configurability of the implementation

Definitions

  • Parity check matrix generation method Parity check matrix generation method, encoding method, communication apparatus, communication system, and encoding
  • the present invention relates to an encoding technique in digital communication, and in particular, a check matrix generation method for generating a parity check matrix for an LDPC (Low-Density Parity Check) code, and the parity check matrix IJ.
  • the present invention relates to an encoding method and a decoding method for encoding predetermined information bits.
  • Non-Patent Document 1 a quasi-cyclic (QC) code (see Non-Patent Document 1) is adopted as an example of an LDPC code
  • the LDPC encoder receives a message (m, m, ⁇ , m) of information length K
  • BP3 ⁇ 4 Binary Phase Shift Keying
  • PSK Quadrature Phase Shift Keying
  • multi-level QAM Quadrature Amplitude Modulation
  • the LDPC decoder in the receiver performs iterative decoding using the “sum-product algorithm” on the demodulated result, and the decoded result (original message m 1, m 2,... ⁇ , M).
  • the parity check matrix for the L DPC code for example, the following non-patent document 1 proposes the following parity check matrix of the QC code (see FIG. 11).
  • I (p) is the row number: r (0 ⁇ r ⁇ p—1), column number: “(r + p ) mod p ”is a cyclic permutation matrix with position“ 1 ”and other positions“ 0 ”.
  • FIG. 12 is a diagram illustrating an example of a parity check matrix represented by a Tanner graph, and corresponds to each column in a parity check matrix H of binary M rows and XN columns of ⁇ 0, 1 ⁇ .
  • the loop is a specific node (in the figure, The inner diameter means the minimum loop, and the length of the loop is expressed by the number of branches that make up the cycle. Accordingly, it is simply expressed as loop 4, loop 6, loop 8.
  • the range of the inner diameter g in the parity check matrix H of the (J, L) QC—LDPC code is “4 ⁇ g ⁇ 12 (g is an even number)” .
  • g 4
  • Non-Patent Document 1 M. Fossorier "Quasi-Cyclic Low Density Parity Check Code IS IT2003, ppl50, Japan, June 29- July 4, 2003.
  • the present invention has been made in view of the above, and generates a parity check matrix for an LDPC code that is compatible with a wide coding rate and has irregular (non-uniform weights for rows and columns).
  • a further object is to obtain a check matrix generation method capable of reducing the circuit scale as compared with the prior art.
  • the check matrix generation method which is the power of the present invention, generates a parity check matrix for LDPC (Low-Density Parity Check) code.
  • a specific cyclic permutation matrix in the regular pseudo-cyclic matrix is converted to a zero matrix, and an irregular masked pseudo-cyclic matrix
  • a non-regular parity check matrix having an LDGM (Low Density Generation Matrix) structure, in which a masking step for generating a signal and a matrix in which the masked pseudo cyclic matrix and the cyclic permutation matrix are arranged in a staircase pattern are arranged at predetermined positions.
  • a parity check matrix generation method comprising: generating a parity check matrix, wherein the mask matrix generation step first corresponds to a reference coding rate (first coding rate) of 1/2 or less.
  • the column degree distribution of the first mask matrix is calculated, and then the second degree corresponding to the second coding rate with the next lowest coding rate is set with the column degree distribution of the first mask matrix as a constraint condition. Calculate the column order distribution of the mask matrix Then, if necessary, the order distribution calculation step for calculating the column order distribution of the third mask matrix, the fourth mask matrix,... In order using the column order distribution of the preceding mask matrix as a constraint.
  • a weight position determining step for determining the weight position of the column of the corresponding mask matrix based on the column degree distribution of the corresponding mask matrix
  • a parity check matrix generated first is a parity check matrix corresponding to the coding rate of 1/2 or less.
  • FIG. 1 is a diagram illustrating a configuration example of a communication system including an LDPC encoder and an LDPC decoder.
  • FIG. 2-1 is a diagram for explaining a masking rule when generating a mask line IJZ corresponding to a code having a coding rate of 1/3.
  • FIG. 2-2 is a diagram for explaining a masking rule when generating a mask line IJZ corresponding to a code with a coding rate of 1/3.
  • FIG. 2-3 is a diagram for explaining a masking rule when generating a mask line IJZ corresponding to a code having a coding rate of 1/3.
  • Fig. 2-4 shows the case of generating a mask row IJZ corresponding to a code with a coding rate of 1/3. It is a figure for demonstrating a masking rule.
  • FIG. 2-5 is a diagram for explaining a masking rule when generating a mask line IJZ corresponding to a code having a coding rate of 1/3.
  • FIG. 3 is a diagram showing a configuration example of an irregular parity check matrix H after masking by a mask matrix Z.
  • FIG. 4 is a diagram illustrating an example of a code configuration method.
  • FIG. 1 A first figure.
  • FIG. 6 is a diagram illustrating an example of codeword generation processing.
  • FIG. 7 is a diagram illustrating an example of codeword generation processing.
  • FIG. 8 is a diagram showing an example of a matrix with a coding rate of 3/4 that is generally disclosed.
  • FIG. 9 shows the case where the matrix shown in FIG. 8 is an M ⁇ N matrix H with a coding rate of 3/4.
  • FIG. 10 is a diagram showing an example of a matrix H ′ of 4M X (7/3) N, where H c is an MX (N—M) matrix obtained by combining arbitrary cyclic permutation matrices. is there.
  • FIG. 11 is a diagram illustrating an example of a parity check matrix of a QC code.
  • FIG. 12 is a diagram illustrating an example of a parity check matrix represented by a Tanner graph. Explanation of symbols
  • FIG. 1 is a diagram illustrating a configuration example of a communication system according to the present embodiment including an LDPC encoder and an LDPC decoder.
  • the transmission side communication device (referred to as a transmission device) includes an LDPC encoder 1 and a modulator 2
  • the reception side communication device (referred to as a reception device) includes a demodulator 4 and LDPC decoding. It is configured to include vessel 5.
  • the parity check matrix generated by the check matrix generation method of the present embodiment that is, M rows subjected to masking processing based on a predetermined masking rule to be described later Generate a parity check matrix H of XN columns.
  • the LDPC encoder 1 receives a message (u 1, u 2,.
  • Generate codeword V of N the information bits are encoded without using the generator matrix G (K: information length, N: codeword length) calculated in the prior art.
  • the codeword V generated by the LDPC encoder 1 is digitally modulated by a predetermined modulation method such as BPSK, QPSK, multi-level QAM, and the like.
  • the signal ⁇ ( ⁇ , X,..., X) is transmitted to the receiving device via the communication path 3.
  • the LDPC decoder 5 in the receiving apparatus performs iterative decoding using a known decoding algorithm, and the decoding result (corresponding to the original message u 1, u 2,.
  • check matrix generation method according to the present embodiment will be described in detail.
  • an irregularity (non-uniform weight distribution) parity check matrix is generated, and that the structure is an LDGM (Low Density Generation Matrix) structure.
  • LDGM Low Density Generation Matrix
  • the check matrix of each embodiment described below The generation process may be executed by the LDPC encoder 1 in the communication device, or may be executed in advance outside the communication device. When it is executed outside the communication device, the generated check matrix is stored in the internal memory.
  • a parity check matrix H of an LDGM structure QC—LDPC code which is a premise of an irregular parity check matrix H after masking processing generated by the check matrix generation processing of this embodiment, is defined. To do.
  • h represents an element of row number m and column number n in the parity check matrix H.
  • I (p) is the row number: r (0 ⁇ r ⁇ p— 1), column number: “(r + p) mod It is a cyclic permutation matrix in which the position of “p” is “1” and the other positions are “0”.
  • 1 (1) can be expressed as the following equation (4 2).
  • the parity check matrix H is a matrix on the left side (part corresponding to information bits).
  • Matrix (part corresponding to parity bit) is a matrix H or H in which I (0) shown in the following formulas (5-1) or (5-2) is arranged in steps. In addition, in the following formulas (5-1) and (5-2)
  • I is the identity matrix and 0 is the zero matrix.
  • the cyclic permutation matrix used in the above step-like structure is not limited to 1 (0), but can be any combination of I (s I se [0, p ⁇ l]). Good.
  • the LDGM structure refers to a structure in which a part of the parity check matrix is a lower triangular matrix, as in the matrix shown in the equation (41).
  • encoding can be easily realized without using the generation matrix G.
  • p in the pseudo-circulant matrix H on the left side of the NORITY check matrix H, p can be
  • 0,32 0,33 0,34 0,35 0,36 (7,87, 36, 15, 53, 11, 16, 136, 59,45, 137, 31, 56,66, 31, 37, 2, 22, 131, 21, 6,92, 56, 72, 53 : 23,21,36,73,127,25,15,10,1,107,18 ⁇
  • the zero matrix is a zero matrix of p rows and Xp columns.
  • the matrix H is the pseudo circulant matrix H
  • Mask matrix Z is masked with 0 elements and the weight distribution is non-uniform, and the distribution of the cyclic permutation matrix of matrix H is the same as the order distribution of mask row ⁇ IJZ
  • the weight distribution of the mask matrix Z when the weight distribution is non-uniform is described later.
  • it is determined by a predetermined density development method.
  • the mask row of 48 rows x 36 mm IJ is expressed as shown in the following formulas (12-1) to (12-5) based on the column order distribution by the density evolution method. .
  • the finally obtained irregular parity check line IJH is, for example, the mask numbers 1 and 48 ( ⁇ number) of 48 / line 36 1 are 0. ⁇ 47) X 36 ⁇ IJ (column number 1 is 0 to 31) pseudo cyclic matrix H, and 48 rows (row number j is 0 to 47) X 48 X ( ⁇ IJ
  • the number 1 can be expressed by the following equation (13) using H of 0 to 47).
  • the parity check matrix ⁇ for generating the LDPC code C is the mask matrix ⁇
  • the masking rule of the present embodiment will be specifically described.
  • a code configuration method up to a coding rate of 3/7 will be described.
  • coding with a coding rate of 3/7 or higher is realized on the premise of the known masking rule described later.
  • FIG. 2-1 is a flowchart showing mask matrix generation processing according to the present embodiment. Before explaining the code configuration method up to coding rate 3/7, refer to Fig. 2-1. A known masking rule will be described. As an example, a mask matrix Z corresponding to a code with a coding rate of 1/3 is generated.
  • H is represented in the above equation (5-2), for example, 32 (the row number is 0 to
  • the matrix is X32 (column number 1 is 0 to 31).
  • LDPC encoder 1 in the communication apparatus sets the size of the mask matrix (step Sl).
  • the mask row ⁇ IJZ corresponding to a code with a coding rate of 1/3 is a matrix of 64 rows X 32 columns
  • the mask matrix Z A corresponding to a code of a coding rate of 1/2 is set to 32 rows X It is a 32-column matrix.
  • the LDPC encoder 1 sets the number of rows of the mask matrix Z A to the maximum order, and uses the mask matrix by the density evolution method with the order distribution of H shown in the above equation (5-2) as a constraint.
  • I H] is an example of a column order distribution, and the numbers (in the columns of column orders 14, 4 and 3 (
  • Fraction represents the column degree distribution of the mask matrix Z A.
  • the column of column degree 2 is the column order of H
  • the LDPC encoder 1 sets the maximum number of rows of the mask matrix Z, and 64 (row number) is 0 to 63 shown in the above equation (51).
  • the column degree distribution of the mask matrix Z is calculated by the density evolution method using the column degree distribution of the mask matrix Z A obtained as described above as a constraint (step S2).
  • the column of column order 3 (4 9 9) / 96 and column order 1 corresponds to the column order distribution of H.
  • the LDPC encoder 1 is a column with a large column order in the mask matrix Z A so that the condition of priority # 1 below is satisfied.
  • the position of “1” in the column of degree 14 (5/64) is determined (step S3). For example, if there is a continuous "1" in Z, there are four places between 1 (0) and a continuous vertical H.
  • step S3 there is only one column with a large column degree of 14, but for example, when there are a plurality of columns with a large column degree, such as 14, 13, 12, ... , to place the columns of the mask matrix Z a left in descending order of column degree (step S3).
  • step S4 LDPC encoder 1, based on the column degree distribution of the mask matrix Z A, so as to satisfy the above priority # 1 condition and following priority # 2 condition, the mask matrix Z A
  • the position of “1” of ⁇ IJ having a small column degree in, that is, ⁇ lJ (18/64, 9/64) of column orders 4 and 3 is determined here (step S4). For example, for a combination of cyclic permutations based on a regular rule, if the arrangement of the mask matrix "1" is also regular, and there is a "1" arrangement of the mask matrix that makes up a particular loop, Because it is regular, there may be multiple identical loops. However, the occurrence probability can be reduced by satisfying the following priority order # 2.
  • Priority # 2 Place based on random numbers
  • the columns of the mask matrix Z A are arranged in descending order of the column order from the left, following the column with the large column order (step S4).
  • the mask matrix Z A generated in steps S3 and S4 is a matrix as shown in equation (14).
  • the LDPC encoder 1 has a large column degree in the mask row ⁇ IJZ so that the condition of the priority order # 1 is satisfied based on the column degree distribution of the mask matrix Z.
  • the position of the column ie, here, the column of the column degree 28 (5/96) is determined as “1” (step S5).
  • "1" may be continuous, but the number should be as small as possible.
  • the LDPC encoder 1 determines that the mask row IJZ satisfies the above-mentioned priority order # 1 condition and the above-mentioned priority order # 2 condition based on the column degree distribution of the mask matrix Z.
  • the position of "1" is determined for a column with a small column order, that is, in this case, some columns (10/96, 8/96, 9/96) of column orders 8, 4, 3 (step S6 ).
  • the columns of the mask matrix Z are arranged in descending order of the column order from the left, following the column with the large column order (step S6).
  • the LDPC encoder 1 avoids performance degradation by increasing the weight of the column having a small column order.
  • Fig. 2-4 is a diagram showing a configuration example of an irregular parity check matrix H after being masked by the mask matrix Z A generated as described above, and Fig. 2-5 is as shown above. Generated
  • FIG. 11 is a diagram illustrating a configuration example of an irregular parity check matrix H after masking with a mask matrix Z.
  • H shown in the figure is the pseudo circulant matrix H.
  • the power that makes the first check matrix a parity check matrix corresponding to coding rate 3/4 is not limited to this, and it corresponds to coding rate 1/2 or less. Any check matrix may be used.
  • an irregularity parity check matrix corresponding to a code with a coding rate of 3/4 is represented by “H”.
  • Z A [Z A XHIH] ".
  • Z A Z A (3/4)
  • H is 12 from the top in the pseudo circulant matrix H (row number j is 0
  • the pseudo cyclic matrix is expressed as ⁇ , ⁇
  • the LDPC encoder 1 of the present embodiment first, in correspondence with the encoding rate of 3/4, the maximum degree the number of rows mask matrix Z A And H
  • the LDPC encoder 1 uses the density evolution method with the order distribution of ⁇ ⁇ (and the column order distribution of the mask matrix ⁇ ⁇ ⁇ obtained above as constraints, corresponding to a coding rate of 3/5.
  • the column degree distribution of the mask matrix ⁇ ⁇ " / ⁇ ' is calculated.
  • the column degree distribution of the previous mask matrix is used as a constraint, and the coding rate is 3/6 by the density evolution method.
  • the column order distribution of the mask matrix is calculated sequentially, and finally the order distribution and sign
  • the column degree distribution of the mask matrix ⁇ ⁇ ⁇ (3/6) corresponding to the coding rate 3/6 is calculated by the density evolution method. To do.
  • the condition of priority # 1 and the condition of priority # 2 are set in the same manner as in the first embodiment. Based on this, the position of “1” in each column of the mask matrix corresponding to the coding rate 3/4, 3/5 is determined, and then the coding rate 3/6, 3 / Determine the position of "1" in each column of the mask matrix corresponding to 7.
  • the mask matrix Z generated by the process of the present embodiment can be expressed as S (12-1) to (12-5).
  • FIG. 3 is a diagram illustrating a configuration example of an irregular parity check matrix H after masking with the mask matrix Z generated as described above. [0090] Next, the case where an LDPC code corresponding to a plurality of coding rates is configured using the irregular parity check matrix H described above will be described.
  • the lower limit of the coding rate in the code generated by the code construction method described above and defined by M and N is set to be between 1/2 and 1/3. Also, in order to realize a coding rate lower than that, good performance is obtained by using repeated transmission described later.
  • FIG. 4 is a diagram showing an example of the code configuration method of the present embodiment.
  • a parity check matrix with a coding rate of 3/4 0 may be inserted into the reception LLR corresponding to the puncturing bit, and normal LDPC decoding may be performed.
  • a novelty is added.
  • decoding is performed using only a partial matrix of the check matrix H corresponding to the coding rate.
  • FIG. 1 A first figure.
  • the coding rate R force is less than 3 ⁇ 4 / 4, that is, the coding rate R is
  • the NORITY bit is punctured in order from the end of the codeword.
  • FIG. 6 is a diagram illustrating the code configuration method of the present embodiment when a coding rate of 3/7 or less is required.
  • the force p described as p in the cyclic permutation matrix being an odd prime number (other than 2) is not limited to this, and an odd number may be selected.
  • the force S which can degrade the performance when puncturing at a high coding rate, is slight.
  • prime numbers it is necessary to have a table and store it beforehand in order to avoid an increase in the amount of calculation, but in the case of an odd number, it is not necessary to store the value.
  • FIG. 8 (“Draft IE Et ⁇ Standard for Local and metropolitan area networks, Part 16: Air interface for Fixed and Mobile Broadband Wireless Access Systems, Amendment for Physical and Medium Access Control Layers for Combined r ixed and Mobile Operation in Licensed Bands ", IEEE P802. 16e / D9 , June 2005), a code rate of 3/7 is realized using a matrix of code rate of 3/4.
  • the numbers in Fig. 8 indicate the value of p, and 1 means the 0 matrix of pXp.
  • a matrix H ′ of M X (7/3) N is constructed as shown in FIG.
  • A is a cyclic permutation matrix
  • the configuration method of the matrix A may be configured by a row IJ by an arbitrary combination of cyclic permutation matrices as in the present embodiment, or a rule as in the first embodiment described above. It may be configured based on specific rules. Further, the process of configuring an LDP C code corresponding to a plurality of coding rates is not limited to the parity check matrix as shown in FIG. 8, and can be applied to any parity check matrix.
  • H is placed on the right side of H and A.
  • the weight distribution of matrix A is derived by the density evolution method.
  • a parity check matrix for LDPC can be generated as in the first embodiment.
  • the code configuration method and the coding method in Embodiment 1 and Embodiment 2 described above can also be applied to coding of an erasure correction code.
  • the parity check matrix generation method and the encoding method according to the present invention are digital It is useful as an encoding technique in wireless communication, and is particularly suitable for communication devices that employ LDPC codes as the encoding method.

Landscapes

  • Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Error Detection And Correction (AREA)

Description

明 細 書
検査行列生成方法、符号化方法、通信装置、通信システムおよび符号化
¾:
技術分野
[0001] 本発明は、ディジタル通信における符号化技術に関するものであり、特に、 LDPC ( Low-Density Parity Check)符号用のパリティ検査行列を生成する検査行列生成 方法、当該ノ リティ検査行歹 IJを用いて所定の情報ビットを符号化する符号化方法、 および復号方法に関するものである。
背景技術
[0002] 以下、符号化方式として LDPC符号を採用する従来の通信システムについて説明 する。ここでは、 LDPC符号の一例として擬似巡回(QC : Quasト Cyclic)符号 (非特許 文献 1参照)を採用する場合について説明する。
[0003] まず、符号化方式として LDPC符号を採用する従来の通信システムにおける、符号 化/復号処理の流れを簡単に説明する。
[0004] 送信側の通信装置 (送信装置と呼ぶ)内の LDPC符号化器では、後述する従来の 方法でパリティ検査行列 Hを生成する。さらに、 LDPC符号化器では、たとえば、 K行 X N列の生成行列G (K :情報長, N :符号語長)を生成する。ただし、 LDPC用のパリ ティ検査行列を H (M行 X N歹 IJ)とした場合、生成行列 Gは、 GHT = 0 (Tは転置行歹 IJ) を満たす行列となる。
[0005] その後、 LDPC符号化器では、情報長 Kのメッセージ (m , m , · · · , m )を受け取り
1 2 K
、このメッセージおよび上記生成行列 Gを用いて、下記(1)式のように、符号語 Cを生 成する。ただし、 H (c, c, …, c ) T = 0とする。
1 2 Ν
C = (m, m, · · ·, m ) G
1 2 K
= (c , c , · · · , c ) …ひ)
1 2 N
[0006] そして、送信装置内の変調器では、 LDPC符号化器で生成した符号語 Cに対して
BP¾ (Binary Phase Shift Keying) , PSK (Quadrature Phase Shift Keying) ,多値 QAM (Quadrature Amplitude Modulation)等の所定の変調方式を用いてデ z変調を行い、その変調信号 X= (X , X , ···, X )を受信装置に送信する。
1 2 N
[0007] 一方、受信側の通信装置 (受信装置と呼ぶ)では、復調器が、受け取った変調信号 y= (y , y , ···, y )に対して、上記 BPSK, QPSK,多値 QAM等の変調方式に応
1 2 N
じたディジタル復調を行い、さらに、受信装置内の LDPC復号器が、復調結果に対し て「sum— productアルゴリズム」による繰り返し復号を実施し、その復号結果(元のメ ッセージ m , m , ···, mに対応)を出力する。
1 2 K
[0008] ここで、 LDPC符号用の従来のパリティ検査行列生成方法を具体的に説明する。 L DPC符号用のパリティ検査行列としては、たとえば、下記非特許文献 1において、以 下の QC符号のパリティ検査行列が提案されて!/、る(図 11参照)。図 11に示す QC符 号のパリティ検査行列は、 5行 X 5列の巡回置換行列(p = 5)が縦方向 (J = 3)と横方 向(L = 5)に配置された行列となっている。
[0009] 一般的には、 M( = pj fXN( = pL^^(J, UQC符号のパリティ検査行列 H
QC
は、下記(2)式のように定義することができる。なお、 pは奇数(2以外)の素数であり、 Lはパリティ検査行列 H における巡回置換行列の横方向(列方向)の個数であり、 J
QC
はパリティ検査行列 H における巡回置換行列の縦方向(行方向)の個数である c
QC
[0010] [数 1]
H QC •(2)
Figure imgf000004_0001
[0011] ただし、 0≤j≤J—l, 0≤1≤L—1において、 I(p )は、行番号: r(0≤r≤p— 1), 列番号:「(r + p )mod p」の位置が" 1 "となり、その他の位置が" 0"となる巡回置換 行列である。
[0012] また、 LDPC符号の設計の際には、一般的に、長さが短いループが多く存在すると きに性能の劣化を引き起こすため、内径を大きくし、長さが短いループ (ループ 4,ル ープ 6等)の数を少なくする必要がある。
[0013] なお、図 12は、検査行列の一例をタナーグラフで表現した場合を示す図であり、 {0 , 1}の 2元の M行 XN列のパリティ検査行列 Hにおいて、各列に対応するノードをビ ットノード b (l≤n≤N)と呼び(図中の〇に相当)、各行に対応するノードをチェックノ n
ード c (l≤m≤M)と呼び(図中の口に相当)、さらに、検査行列の行と列の交点に" m
1"がある場合にそのビットノードとチェックノードを枝で接続する 2部グラフをタナーグ ラフと呼ぶ。また、上記ループとは、図 12に示すように、特定のノード(図中の〇ゃロ に相当)から始まりそのノードで終わる閉路のことを表し、また、内径とは、その最小ル ープを意味する。また、ループの長さは、閉路を構成する枝の数で表現され、長さに 応じて、簡易的にループ 4,ループ 6,ループ 8 · · ·と表現する。
[0014] また、下記非特許文献 1においては、(J, L) QC— LDPC符号のパリティ検査行列 H における内径 gの範囲が、「4≤g≤12 (gは偶数)」とされている。ただし、 g = 4を
QC
回避することは容易であり、多くの場合、 g≥6である。
[0015] 非特許文献1: M.Fossorier "Quasi-Cyclic Low Density Parity Check Code IS IT2003, ppl50, Japan, June 29- July 4, 2003.
発明の開示
発明が解決しょうとする課題
[0016] しかしながら、上記従来の技術によれば、符号化率を変化させるために複数の全く 異なる検査行列が必要であり、それに伴って、メモリ量が大きくなり、回路も複雑にな る、という問題があった。
[0017] 本発明は、上記に鑑みてなされたものであって、広い符号化率に対応可能な、非 正則(行と列の重みが非一様)な LDPC符号用のパリティ検査行列を生成するととも に、さらに、従来技術と比較して回路規模を低減可能な検査行列生成方法を得るこ とを目的とする。
課題を解決するための手段
[0018] 上述した課題を解決し、 目的を達成するために、本発明に力、かる検査行列生成方 法は、 LDPC (Low-Density Parity Check)符号用のパリティ検査行列を生成する 検査行列生成方法であって、巡回置換行列が行方向と列方向に配置されかつ当該 巡回置換行列に特定の規則性を持たせた正則 (行と列の重みが一様)な擬似巡回 行列を生成する擬似巡回行列生成ステップと、前記正則な擬似巡回行列を非正則( 行と列の重みが非一様)にするための、複数の符号化率に対応可能なマスク行列を 生成するマスク行列生成ステップと、特定の符号化率に対応するマスク行列を用いて 、前記正則な擬似巡回行列内の特定の巡回置換行列を 0行列に変換し、非正則な マスク化擬似巡回行列を生成するマスク化ステップと、前記マスク化擬似巡回行列と 巡回置換行列を階段状に配置した行列とを所定位置に配置した、 LDGM (Low De nsity Generation Matrix)構造の非正則なパリティ検査行列を生成する検査行列生 成ステップと、を含む検査行列生成方法であって、前記マスク行列生成ステップでは 、まず、基準となる 1/2以下の符号化率(第 1の符号化率)に対応する第 1のマスク 行列の列次数分布を計算し、つぎに、前記第 1のマスク行列の列次数分布を制約条 件として次に符号化率の低い第 2の符号化率に対応する第 2のマスク行列の列次数 分布を計算し、以降、必要に応じて、前段のマスク行列の列次数分布を制約条件と して第 3のマスク行列、第 4のマスク行列、…の列次数分布を順に計算する次数分布 計算ステップと、符号化率の最も高い第 1のマスク行列から順に、対応するマスク行 列の列次数分布に基づレ、て、当該マスク行列の列の重み位置を決定する重み位置 決定ステップと、を含み、前記検査行列生成ステップでは、最初に生成する検査行 列を、前記符号化率 1/2以下に対応した検査行列とすることを特徴とする。
発明の効果
[0019] この発明によれば、広い符号化率に対応可能な、非正則な LDPC符号用のパリテ ィ検査行列を生成することができる、という効果を奏する。
図面の簡単な説明
[0020] [図 1]図 1は、 LDPC符号化器および LDPC復号器を含む通信システムの構成例を 示す図である。
[図 2-1]図 2— 1は、符号化率 1/3の符号に対応するマスク行歹 IJZを生成する場合の マスキングルールを説明するための図である。
[図 2-2]図 2— 2は、符号化率 1/3の符号に対応するマスク行歹 IJZを生成する場合の マスキングルールを説明するための図である。
[図 2-3]図 2— 3は、符号化率 1/3の符号に対応するマスク行歹 IJZを生成する場合の マスキングルールを説明するための図である。
[図 2-4]図 2— 4は、符号化率 1/3の符号に対応するマスク行歹 IJZを生成する場合の マスキングルールを説明するための図である。
[図 2-5]図 2— 5は、符号化率 1/3の符号に対応するマスク行歹 IJZを生成する場合の マスキングルールを説明するための図である。
[図 3]図 3は、マスク行列 Zによりマスク後のイレギュラーなパリティ検査行列 H の構成 例を示す図である。
[図 4]図 4は、符号構成法の一例を示す図である。
[図 5]図 5は、システムで用意する一番低い符号化率を R = 3/7とした場合の符号を
0
示す図である。
[図 6]図 6は、符号語の生成処理の一例を示す図である。
[図 7]図 7は、符号語の生成処理の一例を示す図である。
[図 8]図 8は、一般に公開されている符号化率 3/4の行列の一例を示す図である。
[図 9]図 9は、図 8に示す行列を符号化率 3/4の M X Nの行列 H とした場合の、 4
CH
M X (7/3) Nの行列 H 'の一例を示す図である。
CH
[図 10]図 10は、任意の巡回置換行列の組み合わせによる M X (N— M)の行列を H c とした場合の、 4M X (7/3) Nの行列 H 'の一例を示す図である。
H CH
[図 11]図 11は、 QC符号のパリティ検査行列の一例を示す図である。
[図 12]図 12は、検査行列の一例をタナーグラフで表現した場合を示す図である。 符号の説明
[0021] 1 LDPC符号化器
2 変!
3 通信路
4 復調器
5 LDPC復号器
発明を実施するための最良の形態
[0022] 以下に、本発明にかかる検査行列生成方法、符号化方法および復号方法の実施 の形態を図面に基づいて詳細に説明する。なお、この実施の形態によりこの発明が 限定されるものではない。
[0023] 実施の形態 1. 図 1は、 LDPC符号化器および LDPC復号器を含む本実施の形態の通信システム の構成例を示す図である。図 1において、送信側の通信装置 (送信装置と呼ぶ)は、 LDPC符号化器 1と変調器 2を含む構成とし、受信側の通信装置 (受信装置と呼ぶ) は、復調器 4と LDPC復号器 5を含む構成とする。
[0024] ここで、 LDPC符号を採用する通信システムにおける符号化処理,復号処理の流 れを簡単に説明する。
[0025] 送信装置内の LDPC符号化器 1では、本実施の形態の検査行列生成方法により 生成されたパリティ検査行列、すなわち、後述する所定のマスキングルールに基づい てマスキング処理が行われた M行 X N列のパリティ検査行列 Hを生成する。
[0026] その後、 LDPC符号化器 1では、情報長 Kのメッセージ (u , u , ···, u )を受け取り
1 2 K
、このメッセージおよび上記ノ リティ検査行列 Hを用いて、下記(3)式のように、長さ
Nの符号語 Vを生成する。なお、本実施の形態においては、従来技術において計算 していた生成行列 G(K:情報長, N:符号語長)を用いずに、情報ビットの符号化処 理を行う。
ν= { (ν , V , ·· ·, V )≡GF(2) (v , v , ·· ·, v )H τ = 0}
1 2 N I 1 2 N M
-• (3)
[0027] そして、送信装置内の変調器 2では、 LDPC符号化器 1で生成した符号語 Vに対し て、 BPSK, QPSK,多値 QAM等の所定の変調方式によりディジタル変調を行い、 その変調信号 χ= (χ , X , ·· ·, X )を、通信路 3を介して受信装置に送信する。
1 2 N
[0028] 一方、受信装置では、復調器 4が、通信路 3を介して受け取った変調信号 y= (y ,
1 y,…, y )に対して、上記 BPSK, QPSK,多値 QAM等の変調方式に応じたデイジ
2 N
タル復調を行い、さらに、受信装置内の LDPC復号器 5が、公知の復号アルゴリズム による繰り返し復号を実施し、その復号結果(元のメッセージ u , u , ···, uに対応)を
1 2 K 出力する。
[0029] つづいて、本実施の形態における検査行列生成方法を詳細に説明する。なお、本 実施の形態においては、イレギュラー(重み分布が非一様)なノ リティ検査行列を生 成することとし、その構造として、 LDGM (Low Density Generation Matrix)構造を 採用することを前提とする。また、以下において説明する各実施の形態の検査行列 生成処理については、通信装置内の LDPC符号化器 1で実行することとしてもよいし 、または、通信装置の外部で予め実行しておくこととしてもよい。通信装置の外部で 実行される場合には、生成済みの検査行列を内部メモリに記憶しておく。
[0030] まず、本実施の形態の検査行列生成処理により生成されるマスキング処理後のィレ ギュラーなパリティ検査行列 H の前提となる、 LDGM構造の QC— LDPC符号のパ リティ検査行列 H を定義する。
QCL
[0031] たとえば、 M ( = pJ)行 X N ( = pL + pJ)列の LDGM構造の QC— LDPC符号のパ リティ検査行列 H (= [h ] )は、下記 (4 1)式のように定義することができる。
QCL
[0032] [数 2-1]
Figure imgf000009_0001
1 )
[0033] なお、 h は、パリティ検査行列 H において、行番号 m,列番号 nの要素を表す。
QCL
また、 0≤j≤J— 1 , 0≤1≤L— 1において、 I (p )は、行番号: r (0≤r≤p— 1) ,列番 号:「(r + p ) mod p」の位置が" 1 "となり、その他の位置が" 0"となる巡回置換行列 である。たとえば、 1 (1)は、下記(4 2)式のように表すことができる。
[0034] [数 2-2]
Figure imgf000009_0002
上記パリティ検査行列 H は、左側の行列(情報ビットに対応する部分)力 上記(
QCL
2)式で示した QC符号のパリティ検査行列と同一の擬似巡回行列 H であり、右側の 行列 (パリティビットに対応する部分) 、下記(5— 1)式または(5— 2)式に示す I (0) を階段状に配置した行列 Hまたは Hである。なお、下記(5— 1)式, (5— 2)式にお
T D
いて、 Iは単位行列であり、 0は零行列である。
[0036] [数 3-1]
0 0 0
I I 0 0
■(5-1)
0 I I 0
0 0 I I
[0037] [数 3-2]
/(0) 0
/(0) /(0)
0 '·. .(5—2)
/(0) 7(0) 0
0 /(0) /(0)
[0038] ただし、上記階段状の構造で用いている巡回置換行列は、 1(0)に限定しているわ けではなぐ任意の I(s I se[0, p—l])の組み合わせでもよい。
[0039] また、上記 LDGM構造とは、上記 (4 1)式に示す行列のように、パリティ検査行 列の一部を下三角行列にした構造のことをいう。この構造を用いることにより、生成行 列 Gを用いずに符号化を容易に実現できる。たとえば、組織符号語 Vを下記(6)式の ように表し、情報メッセージ u= (u , u , ···, u )が与えられた場合、パリティ要素 p =
1 2 K m
(p , p , ···, p )は、「Η ·ντ = 0」を満たすように、すなわち、下記(7)式のように生
1 2
成する
(6)
Figure imgf000010_0001
1 :
ただし、 N二 K + Mである。
[0041] [数 4]
Figure imgf000010_0002
\<m≤M
\<n<N ■(7) [0042] さらに、本実施の形態においては、上記(4 1)式のように定義された LDGM構造 の QC— LDPC符号のパリティ検査行列 H に、特定の規則性を設けている。具体
QCL
的には、ノ リティ検査行列 H の左側の擬似巡回行列 H 部分において、 p を任意
QCL QC 0,1 の整数とした場合、行番号 j( = 0, 1, 2, --]-1),列番号 1(=0, 1, 2, …し— 1)に 配置された ρ行 Xp列の巡回置換行列 Ι(ρ )の に、下記(8— 1)式、(8— 2)式、(8 3)式または(8— 4)式となる規則性を設ける。
[0043] p =p (j + l)mod p
j,l 0,1
-•(8-1)
p =((p— p ) (j + l))mod p
j,l 0,1
-•(8-2)
p = (((p — p ) (j + l))mod p )mod p
j,l A 0,1 A
p =157
A
o≤j≤j-i
p: prime number
-•(8-3)
p = ( (p (j +丄" mod p )mod p
j,l 0,1 A
p =157
A
o≤j≤j-i
p: prime number
-•(8-4)
[0044] なお、上記 pは、システムが要求する MAXの情報長の時に最も良い性能が得られ
A
るように決定する。上記「p =157」は、情報長 pXL( = 36)カ 5000〜6000近辺を
A
想定した場合の値となる。
[0045] たとえば、上記(8— 4)式を用い、 L = 36とした場合、一例として下記(9)式のように 表すことができる。
{p , P , P , P , P , P , P , P , P , P , P ; , P , P , P , P , P
0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 0,10 0,11 0,12 0,13 0,14 0,15 0,16
P , P , P : , P , P , P , P , P , P , P , P , P , P , P , P
0,17 0,18 0,19 0,20 0,21 0,22 0,23 0,24 0,25 0,26 0,27 0,28 0,29 0,30 0,31
P , P , P : , P , P }
0,32 0,33 0,34 0,35 0,36 = {7,87, 36, 15, 53, 11, 16, 136, 59,45, 137, 31, 56,66, 31, 37, 2, 22, 131, 21, 6,92, 56, 72, 53: 23,21,36,73,127,25,15,10,1,107,18}
-• (9)
[0046] つづいて、本実施の形態の検査行列生成方法における特徴的な処理である、ノ リ ティ検査行列 H に対するマスク処理につ!/、て説明する。
QCL
[0047] たとえば、上記 (4 1)式に示す左側の行列を、下記(10— 1)式に示すように、 J X Lの擬似巡回行列 H と表し、マスク行列 Z(=[z ])を GF (2)上の J行 XL列の行列
QC
とした場合、後述する所定のルールを適用すると、マスク処理後の行列 H
MQC
記(10— 2)式のように表すことができる c
[0048] [数 5-1]
Figure imgf000012_0001
[0049] [数 5-2]
Figure imgf000012_0002
[0050] なお、上記(10— 2)式における z I(p )は、下記(11)式のように定義される c
[0051] [数 6]
I(Pjj) for ζ,.; = 1
ノ ( ,) = ■(11)
0 for . -0
[0052] 上記 0行列は、 p行 Xp列の 0行列である。また、行列 H は、擬似巡回行列 H を
MQC QC
マスク行列 Zの 0要素によりマスクし、重み分布を非一様 (イレギュラー)にした行列で あり、また、行列 H の巡回置換行列の分布は、マスク行歹 IJZの次数分布と同じであ
MQC
[0053] ただし、上記重み分布を非一様にする場合のマスク行列 Zの重み分布は、後述す るように、所定の密度発展法で求めることとする。たとえば、 48行 X 36歹 IJのマスク行 列は、密度発展法による列次数分布に基づレ、て、下記( 12— 1 )式〜( 12— 5)式の ように表すことカでさる。
[数 7-1]
Figure imgf000013_0001
[0055] [数 7-2]
Figure imgf000013_0002
■(12-2)
[0056] [数 7-3]
Figure imgf000013_0003
■(12 - 3)
[0057] [数 7-4]
Figure imgf000014_0001
■(12-4)
[0058] [数 7-5]
1 0 0 0 0
0 0 0 0 0 0 0 0 0 1 0 0 0 0
0 0 0 0
1 1 0 0 0 0 0 0 0 0 0
ZD= 0 0
1 1
0 0 0 0
1 0 0 0 0 0
0 1 0 0
1 0 0 0 0 0 0 0 1
0 0 0 0 0 0 0
■(12-5)
[0059] したがって、本実施の形態において、最終的に求めるイレギュラーなパリティ検查行 歹 IJH は、たとえば、 48/行 36歹1】のマスクネ亍歹1】 、 48亍(ί 番号】は 0〜47) X 36歹 IJ ( 列番号 1は 0〜31 )の擬似巡回行列 H 、および 48行(行番号 jは 0〜47) X 48歹リ(歹 IJ
QC
番号 1は 0〜47)の Hを用いて、下記(13)式のように表すことができる。
T
[0060] H = [Z X H
M QC I H ]
T
= [H I Η ] - - - (13)
MQC T
[0061] すなわち、 LDPC符号 Cを生成するためのパリティ検査行列 Η は、マスク行列 Ζ
MQC
と擬似巡回行列 Η の行番号 j = 0の巡回置換行列の値の設計により、与えられる。
QC
[0062] つづいて、本実施の形態のマスキングルールを具体的に説明する。本実施の形態 においては、一例として、符号化率 3/7までの符号構成法について説明する。ここ では、後述する既知のマスキングルールを前提として、符号化率 3/7以上の符号化 を実現する。
[0063] 図 2—1は、本実施の形態のマスク行列生成処理を示すフローチャートである。ここ で、符号化率 3/7までの符号構成法について説明する前に、図 2— 1にしたがって 、既知のマスキングルールについて説明する。一例として、符号化率 1/3の符号に 対応するマスク行列 Zを生成する。
[0064] なお、ここでは、マスク行列生成処理を、通信装置内の LDPC符号化部 1にて実行 する場合の例を示す。また、符号化率 1/2の符号に対応するイレギュラーなパリティ 検査行列 H を「H =[ZAXH I H ]」と仮定する。ただし、 H は、上記擬
M(l/2) M(l/2) QCU D QCU
似巡回行列 H における上半分の 32(行番号 jは 0〜31) X32(列番号 1は 0〜31)
QC
の擬似巡回行列を表し、 Hは、上記(5— 2)式に示す、たとえば、 32(行番号は0〜
D
31) X32(列番号 1は 0〜31)の行列とする。
[0065] まず、通信装置内の LDPC符号化器 1では、マスク行列のサイズを設定する(ステツ プ Sl)。ここでは、一例として、符号化率 1/3の符号に対応するマスク行歹 IJZを 64行 X 32列の行列とし、符号化率 1/2の符号に対応するマスク行列 ZAを 32行 X 32列 の行列とする。
[0066] つぎに、 LDPC符号化器 1は、マスク行列 ZAの行数を最大次数とし、上記(5— 2) 式に示す Hの次数分布を制約条件として、密度発展法により、マスク行列 ZAの列次
D
数分布を計算する(ステップ S2)。図 2— 2は、パリティ検査行列 H (=[ZAXH
M(l/2) QCU
I H ])の列次数分布の一例を示す図であり、列次数 14, 4, 3の欄に記載の数字(
D
分数)がマスク行列 ZAの列次数分布を表している。なお、列次数 2の欄は、 H の列次
D
数分布に対応している。
[0067] さらに、 LDPC符号化器 1は、マスク行列 Zの行数を最大次数とし、上記(5 1)式 に示す 64(行番号】は 0〜63) X64(歹 IJ番号 1は 0〜63)の Hの次数分布、および上
τ
記で求めたマスク行列 ZAの列次数分布、を制約条件として、密度発展法により、マス ク行列 Zの列次数分布を計算する (ステップ S2)。図 2— 3は、符号化率 1/3の符号 に対応するイレギュラーなパリティ検査行列 H (=[ZXH | H ])の列次数分 布の一例を示す図であり、列次数 28, 8, 4, 3 (9列分)の欄に記載の数字がマスク行 列 Zの列次数分布を表している。なお、列次数 3の(41 9) /96と列次数 1の欄は、 Hの列次数分布に対応している。
T
つぎに、 LDPC符号化器 1は、上記マスク行列 ZAの列次数分布に基づいて、下記 優先順位 # 1の条件を満たすように、マスク行列 ZAにおける列次数の大きい列、すな わち、ここでは列次数 14の列(5/64)の、 "1"の位置を決める(ステップ S3)。たとえ ば、 Zに連続する" 1"が存在する場合は、 Hの縦に連続する 1 (0)との間で 4箇所に
D
巡回置換行列が存在し、それらがループ 4を構成する可能性があるため、下記優先 順位 # 1の条件を満たすことにより、その可能性を排除する。
[0069] ·優先順位 # 1の条件:同一列内の" 1 "の間隔は 2行以上離す
[0070] このとき、上記列次数の大きい列は、 "1"の密度が大きいため、後述する優先順位
# 2の条件を満たす必要はない。また、上記で求めた列次数 14を実現するために、 同一列において "1"が連続しても良いが、その数は可能な限り少なくする。
[0071] そして、ここでは、列次数の大きい列が 14の一つしか存在していないが、たとえば、 列次数の大きい列が 14, 13, 12· · ·等のように複数存在する場合は、左から列次数 の降順にマスク行列 ZAの列を配置する(ステップ S3)。
[0072] つぎに、 LDPC符号化器 1は、上記マスク行列 ZAの列次数分布に基づいて、上記 優先順位 # 1の条件および下記優先順位 # 2の条件を満たすように、マスク行列 ZA における列次数の小さい歹 IJ、すなわち、ここでは列次数 4, 3の歹 lJ (18/64, 9/64) の、 "1"の位置を決める(ステップ S4)。たとえば、規則的なルールに基づく巡回置換 行列の組み合わせに対して、マスク行列の" 1 "の配置も規則的である場合、特定の ループを構成するマスク行列の" 1 "の配置が存在すると、規則的であるがゆえに多 重的に同一ループが多数個存在する可能性がある。し力、しながら、下記優先順位 # 2の条件を満たすことにより、その発生確率を減らすことができる。
[0073] ,優先順位 # 2の条件:乱数に基づき配置する
[0074] そして、上記列次数の大きい列につづいて、左から列次数の降順にマスク行列 ZA の列を配置する (ステップ S4)。
[0075] たとえば、上記ステップ S3および S4にて生成されたマスク行列 ZAは、(14)式に示 すような行列となる。
[0076] [数 8] 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 o"o
0 1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0
1 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0
1 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 0 0
0 1 0 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0
0 0 1 0 1 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0
1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0
1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0
0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
1 0 1 1 1 0 0 0 0 0 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0
1 1 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0
1 0 0 t 0 0 0 0 \ 0 0 0 0 0 0 0 1 0 0 0 t 0 0 0 1 0 0 0 0 0 0 0
0 1 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
0 0 1 1 1 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 0 1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0
0 1 0 1 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1
0 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
1 0 0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0
0 1 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
0 1 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0
1 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0
0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0
0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0
1 0 0 0 0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
0 1 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0
0 1 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1 0 0
0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
1 1 0 0 1 0 0 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
Lo 1 0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0」
[0077] なお、密度発展法により求めたマスク行列 ZAの列次数分布に基づ!/、てマスク行列 Z Aを構成する場合、小さい列次数の列(列次数 3, 4等)に小さいループ (ループ 4, 6 等)が含まれていると、誤り確率が上がり、良好な性能が得られない場合がある。この ような場合、 LDPC符号化器 1は、小さい列次数の列の重みを増やすことで、性能劣 化を回避する。上記マスク行列 ZAの場合は、図 2— 2に示す列次数分布において、 列次数 3の列(9/64)のうちの 3列の重みを増やしている。すなわち、上記(14)式 においては、列次数 14が 5列,列次数 4が 21列,列次数 3が 6列となっている。
[0078] つぎに、 LDPC符号化器 1は、上記マスク行列 Zの列次数分布に基づレ、て、上記優 先順位 # 1の条件を満たすように、マスク行歹 IJZにおける列次数の大きい列、すなわ ち、ここでは列次数 28の列(5/96)の、 "1"の位置を決める(ステップ S5)。このとき 、上記列次数の大きい列は、 " の密度が大きいため、上記優先順位 # 2の条件を 満たす必要はない。また、上記で求めた列次数 28を実現するために、同一列におい て "1"が連続しても良いが、その数は可能な限り少なくする。
[0079] そして、ここでは、列次数の大きい列が 28の一つしか存在していないが、列次数の 大きレ、列が複数存在する場合は、左から列次数の降順にマスク行列 Zの列を配置す る(ステップ S 5)。 [0080] つぎに、 LDPC符号化器 1は、上記マスク行列 Zの列次数分布に基づいて、上記優 先順位 # 1の条件および上記優先順位 # 2の条件を満たすように、マスク行歹 IJZにお ける列次数の小さい列、すなわち、ここでは列次数 8, 4, 3の一部の列(10/96, 8 /96, 9/96)の、 "1 "の位置を決める(ステップ S6)。
[0081] そして、上記列次数の大きい列につづいて、左から列次数の降順にマスク行列 Zの 列を配置する(ステップ S6)。
[0082] なお、密度発展法により求めたマスク行列 Zの列次数分布に基づいてマスク行列 Z を構成する場合、小さい列次数の列(列次数 3, 4, 8等)に小さいループ (ループ 4, 6等)が含まれていると、誤り確率が上がり、良好な性能が得られない場合がある。こ のような場合、 LDPC符号化器 1は、小さい列次数の列の重みを増やすことで、性能 劣化を回避する。
[0083] また、 LDPC符号を多値変調のような不均一な誤り確率のアプリケーションに適用 する場合、誤り確率の小さ!/、ビットを列次数の大き!/ヽ列に対応するビットに割り当て、 誤り確率の大き!/、ビットを列次数の小さ!/、列に対応するビットに割り当てると、性能が 向上する。このような場合に、上記ステップ S3〜S6に示すように、左側から列次数の 降順に列を並べておくと、ビットのオーダリングが容易となる。
[0084] 図 2— 4は、上記のように生成したマスク行列 ZAによりマスク後の、イレギュラーなパ リティ検査行列 H の構成例を示す図であり、図 2— 5は、上記のように生成したマ
M(l/2)
スク行列 Zによりマスク後の、イレギュラーなパリティ検査行列 H の構成例を示す 図である。なお、図 2— 5において、図示の H は、上記擬似巡回行列 H における
QCD QC
下半分の 32 (行番号 jは 32〜63) X 32 (列番号 1は 0〜31)の擬似巡回行列を表す。 また、図示の 0は 0行列を表し、図示の Iは単位行列を表し、 Hと Iと 0で Hを構成する
[0085] つぎに、上記符号化率 1/3の符号生成処理を前提として、符号化率 3/7以上の 符号化を実現する方法にっレ、て説明する。本実施の形態のマスク行列生成処理とし て、上記符号化率 1/3の符号生成処理と異なる処理について説明する。なお、本 実施の形態においては、後述するように、最初に構成する検査行列を符号化率 3/ 4に対応した検査行列としている力 これに限らず、符号化率 1/2以下に対応する 検査行列であればよい。
[0086] たとえば、符号化率 3/4の符号に対応するイレギュラーなノ リティ検査行列を「H
M(
= [ZAX H I H ]」と仮定する。ただし、 ZA( = ZA(3/4))は l2行 X 36列のマス ク行列であり、 H は、擬似巡回行列 H における上から 1/4の 12 (行番号 jは 0
QC(3/4) QC
〜11) X 36 (列番号 1は 0〜35)の擬似巡回行列を表し((4 1)式参照)、 H は
TO/4) 上記 Hである((5— 2)式参照)。また、符号化率 3/5, 3/6 = 1/2, 3/7に対応
D
したイレギュラーなパリティ検査行列を H , H , H ( = H )と表し、マスク行 列をそれぞれ ΖΛ , ΖΛ , ΖΛ""' ( = Ζ)と表し、擬似巡回行列をそれぞれ Η , Η
, Η ( = Η )と表し、 1 (0)を階段状に配置した行列をそれぞれ Η , Η
QC(l/2) QC(3/7) QC T(3/5) T(L
, Η (=Η )と表す。
2) Τ(3/7) Τ
[0087] たとえば、図 2 1のステップ S2の処理において、本実施の形態の LDPC符号化 器 1では、まず、符号化率 3/4に対応して、マスク行列 ZAの行数を最大次数とし、 H
T
(=τ )の次数分布を制約条件として、密度発展法により、マスク行列 ζΛの列次数 分布を計算する。つぎに、 LDPC符号化器 1は、符号化率 3/5に対応して、上記 Η τ( の次数分布および上記で求めたマスク行列 ΖΑの列次数分布を制約条件として、 密度発展法により、マスク行列 ζΛ"'の列次数分布を計算する。つぎに、一つ前のマ スク行列の列次数分布を制約条件として、密度発展法により、符号化率 3/6に対応 したマスク行列の列次数分布を順次計算し、最後に、上記 Ηの次数分布および符
Τ
号化率 3/6に対応したマスク行列 ΖΑ(3/6)の列次数分布を制約条件として、密度発展 法により、符号化率 3/7に対応したマスク行列 Ζの列次数分布を計算する。
[0088] つぎに、 LDPC符号化器 1では、図 2—1のステップ S3〜S6の処理において、前述 した実施の形態 1と同様に、優先順位 # 1の条件および優先順位 # 2の条件に基づ いて、符号化率 3/4, 3/5に対応するマスク行列の各列の" 1"の位置を決め、その 後、さらに、同様の手順で、符号化率 3/6, 3/7に対応するマスク行列の各列の" 1 "の位置を決める。上記本実施の形態の処理にて生成されたマスク行列 Zは、上記( 12— 1)式〜(12— 5)式のように表すこと力 Sできる。
[0089] 図 3は、上記のように生成したマスク行列 Zによりマスク後の、イレギュラーなパリティ 検査行列 H の構成例を示す図である。 [0090] つづいて、前述のイレギュラーなパリティ検査行列 H を用いて、複数の符号化率に 対応した LDPC符号を構成する場合について説明する。
[0091] たとえば、前述の符号構成法によって生成され、 Mと Nで規定された符号における 符号化率の下限を 1/2から 1/3の間とした。また、それ以下の符号化率を実現する ためには、後述する繰り返し送信を用いて良好な性能を得る。
[0092] また、図 4は、本実施の形態の符号構成法の一例を示す図であり、たとえば、符号 化率 0. 75の符号を基準符号語とし、それよりも高い符号化率( = 0. 9)の符号語を 生成する場合はパリティのパンクチヤを行う。なお、この際の復号には符号化率 3/4 の検査行列を使って、パンクチャリングビットに対応する受信 LLRに 0を揷入し、通常 の LDPC復号を行えばよい。一方、低い符号化率( = 3/5)の符号語を生成する場 合は、ノ リティを追加している。この際の復号には、図 3に示すように、符号化率に対 応した検査行列 H の部分行列のみを用いて復号する。
[0093] ここで、複数の符号化率に対応した LDPC符号の構成法を具体的に説明する。た とえば、システムで用意する一番低い符号化率を R = 1/2以下とする。図 5は、たと
0
えば、システムで用意する一番低い符号化率を R = 3/7とした場合の符号を示す
0
図である。
[0094] たとえば、符号化率 R = 3/7に対応する符号がメモリに記憶され、符号化率 Rの
0 1 符号を構成する場合、符号化率 R力 ¾/4未満であれば、すなわち、符号化率 Rが
1 1
3/4〜3/7の間であれば、ノ リティビットを符号語の最後尾から順にパンクチヤする
[0095] 一方で、たとえば、 3/7以下の符号化率が必要となった場合、本実施の形態では 、図 6に示すように、情報長 K,符号長 Nの符号化率 K/N = 3/7の符号語(図 6中 の A+ノ リティビット)に、列重みの重い順から選択した符号語ビット B (長さ b:各列に 対応するビットの情報は Aと同じ)を加えて符号化率 K/ (N + b)の符号語を生成す る。たとえば、 b = (2/3) * Kの場合、符号化率は 1/3となり、 b = Kの場合は、符号 化率 3/10となる。図 6は、 3/7以下の符号化率が必要となった場合における、本 実施の形態の符号構成法を示す図である。
[0096] また、 b= (2/3) * K (図 7の Bの箇所)のように同一の符号語の一部が繰り返し 2 回にわたって送られ、さらにそれ以上に低い符号化率が必要な場合は、再度「A+ ノ リティビット + B」の符号語ビットを繰り返し送ることも可能とした。これは、列重みの 重い Bの箇所のみ重複して送ることにより性能が向上する事が確認されているためで ある。また、さらに低い符号化率を実現するためには、この部分 (A +ノ リティビット + B)をさらに重複して送る。
[0097] そして、上記の方法により符号化された符号語が通信路を通って受信機で受信さ れ、復号器が誤りを訂正する際に符号語の一部あるいは全てが重複した場合には、 その重複したビットの受信値を、重複した個数分だけ加算平均し、その結果を復号器 に渡すこととした。また、列重みの重い列に対応する符号語ビットは復号性能がよくな ること力 S知られている。したがって、本実施の形態では、上記処理によりノイズ成分の 分散値を下げることができ、それに伴って信頼度が上がる(対応するビットの誤り確率 が下がり)ため、復号性能を向上させることができる。
[0098] なお、本実施の形態において、巡回置換行列の pが奇数(2以外)の素数であること として記載した力 pについてはこれに限らず、奇数を選んでもよい。この場合、高符 号化率のパンクチヤ時に性能が劣化する可能性がある力 S、劣化は僅かである。一方 で、素数については、計算量が増加することを回避するためにテーブルを持ち予め 記憶しておく必要があるが、奇数の場合はその値を記憶しておく必要がない。
[0099] このように、本実施の形態においては、従来技術とは異なり、符号化率を変化させ るために複数の全く異なる検査行列をメモリに記憶しておく必要がないため、従来よ りもメモリ量を小さくでき、回路を簡単化できる。
[0100] また、本実施の形態においては、たとえば、符号化率 1/2に対応する検査行列を 基準に検査行列を拡大しているときと比べて、符号化率 3/4に対応するより小さい 検査行列から拡大することとし、また、 3/4以上の符号化率の際には符号化率 3/4 に対応する検査行列に従って復号処理を行うこととしたので、符号化率 1/2に対応 する検査行列を基準にする場合と比較して復号に関する計算量を低減することがで きる。また、 3/7以下の符号化率であっても、列次数の重い箇所の符号語の再送回 数を多くすることによって、低い符号化率を実現することができるので、本実施の形態 においては、性能の劣化を抑えながら、より広い符号化率を実現することができる。 [0101] 実施の形態 2.
前述の実施の形態 1では、巡回置換行列 I (p )を一定のルールで構成した例につ いて示した力 本実施の形態では、たとえば、一般に公開されている図 8 ("Draft IE Et^ Standard forLocal and metropolitan area networks, Part 16 : Air interfac e for Fixed and Mobile Broadband Wireless Access Systems, Amendment fo r Physical and Medium Access Control Layers for Combined r ixed and Mo bile Operation in Licensed Bands" , IEEE P802. 16e/D9, June 2005)に示すよ うな符号化率 3/4の行列を用いて、符号化率 3/7を実現する。なお、図 8中の数字 は p の値を示しており、 1は p X pの 0行列を意味する。
[0102] 具体的には、図 8に示す行列を符号化率 3/4の M X Nの行列 H とした場合、 4
CH
M X ( 7/3) Nの行列 H 'を図 9のように構成する。たとえば、 Aは巡回置換行列の
CH
組み合わせによる行列で構成され、さらに、単位行列 Iを 2つ連結させて右下に階段 状に配置する。行列 Aの重み分布は密度発展法で導出する。この方法により、前述 した実施の形態 1と同様に LDPC用のパリティ検査行列を生成することができる。
[0103] なお、行列 Aの構成法は、本実施の形態のように巡回置換行列の任意の組み合わ せによる行歹 IJで構成することとしてもよいし、前述した実施の形態 1のように規則的な ルールに基づいて構成することとしてもよい。また、複数の符号化率に対応した LDP C符号を構成する処理については、上記図 8に示すようなパリティ検査行列に限らず 、どのようなパリティ検査行列においても適用可能である。
[0104] また、たとえば、任意の巡回置換行列の組み合わせによる M X (N - M)の行列を H とする場合には、 4M X ( 7/3) Nの行列 H 'を図 10のように構成する。 Aは巡
CH CH
回置換行列の組み合わせによる行列で構成され、さらに、 H と Aの右側に Hを配
CH D
置する。行列 Aの重み分布は密度発展法で導出する。この方法により、前述の実施 の形態 1と同様に、 LDPC用のノ リティ検査行列を生成することができる。
[0105] また、前述した実施の形態 1および上記実施の形態 2における符号構成法および 符号化法は、消失訂正符号の符号化にも適用可能である。
産業上の利用可能性
[0106] 以上のように、本発明にかかる検査行列生成方法および符号化方法は、ディジタ ル通信における符号化技術として有用であり、特に、符号化方式として LDPC符号を 採用する通信装置に適して!/、る。

Claims

請求の範囲
[1] LDPC (Low-Density Parity Check)符号用のパリティ検査行列を生成する検査 行列生成方法において、
巡回置換行列が行方向と列方向に配置されかつ当該巡回置換行列に特定の規則 性を持たせた正則(行と列の重みが一様)な擬似巡回行列を生成する擬似巡回行列 生成ステップと、
前記正則な擬似巡回行列を非正則 (行と列の重みが非一様)にするための、複数 の符号化率に対応可能なマスク行列を生成するマスク行列生成ステップと、 特定の符号化率に対応するマスク行列を用いて、前記正則な擬似巡回行列内の 特定の巡回置換行列を 0行列に変換し、非正則なマスク化擬似巡回行列を生成する マスク化ステップと、
前記マスク化擬似巡回行列と巡回置換行列を階段状に配置した行列とを所定位置 に配置した、 LDGM (Low Density Generation Matrix)構造の非正則なパリティ検 查行列を生成する検査行列生成ステップと、
を含む検査行列生成方法であって、
前記マスク行列生成ステップでは、
まず、基準となる 1/2以下の符号化率(第 1の符号化率)に対応する第 1のマスク 行列の列次数分布を計算し、つぎに、前記第 1のマスク行列の列次数分布を制約条 件として次に符号化率の低い第 2の符号化率に対応する第 2のマスク行列の列次数 分布を計算し、以降、必要に応じて、前段のマスク行列の列次数分布を制約条件と して第 3のマスク行列、第 4のマスク行列、…の列次数分布を順に計算する次数分布 十昇 アツノど、
符号化率の最も高!/、第 1のマスク行列から順に、対応するマスク行列の列次数分布 に基づいて、当該マスク行列の列の重み位置を決定する重み位置決定ステップと、 を含み、
前記検査行列生成ステップでは、最初に生成する検査行列を、前記符号化率 1/ 2以下に対応した検査行列とすることを特徴とする検査行列生成方法。
[2] 前記検査行列生成ステップでは、最後に生成する検査行列を、符号化率 1/3未 満に対応する検査行列とすることを特徴とする請求項 1に記載の検査行列生成方法
[3] 前記重み位置決定ステップでは、
列次数が小さいことに起因して発生する誤りの確率に基づいて、マスク行列生成処 理を、列次数の大きレ、列と小さレ、列に分けて行うこととし、
列次数の小さい列については、「同一列内の重みの間隔を 2行以上離す」という第 1の条件および「乱数に基づいて配置する」という第 2の条件を満たすように、マスク 行列の列の重み位置を決定することを特徴とする請求項 2に記載の検査行列生成方 法。
[4] 前記重み位置決定ステップでは、前記第 2の条件を満たすように、
所定の方法で乱数列を生成し、
当該乱数列において差分が 1の要素が存在する場合に、一方の要素を乱数列の 最後に移動することにより、要素間の差分力 ¾以上となる擬似乱数系列を生成し、 当該擬似乱数系列を列次数毎に分割して、それぞれを列重みの行位置番号とする ことを特徴とする請求項 3に記載の検査行列生成方法。
[5] 前記重み位置決定ステップでは、列次数分布における列次数の降順に、マスク行 列の列を配置することを特徴とする請求項 3に記載の検査行列生成方法。
[6] 前記重み位置決定ステップでは、列次数分布における列次数の降順に、マスク行 列の列を配置することを特徴とする請求項 4に記載の検査行列生成方法。
[7] LDPC (Low-Density Parity Check)符号用のパリティ検査行列を用いて所定の 情報ビットを符号化する符号化方法であって、
前記請求項 2に記載の処理で生成された非正則なパリティ検査行列を用いて所定 の情報ビットを符号化する符号化ステップ、
を含むことを特徴とする符号化方法。
[8] 前記符号化ステップでは、情報長 K,符号長 N (情報ビット A+ノ リティビット P)の符 号化率 K/Nの符号語に、情報ビット Aのうち列重みの重い順から選択した符号語ビ ット B (ビット長 b)を加えて、符号化率 K/ (N + b)の符号語「A+P + B」を生成するこ とを特徴とする請求項 7に記載の符号化方法。
[9] 前記符号化率 K/Nを 3/7とし、前記符号化率 K/ (N + b)を 1/3とすることを特 徴とすることを特徴とする請求項 8に記載の符号化方法。
[10] さらに低い符号化率が必要な場合は、前記「A+ P + B」を繰り返した符号語ビット を生成することを特徴とする請求項 9に記載の符号化方法。
[11] LDPC (Low-Density Parity Check)符号用のパリティ検査行列を生成する通信 装置であって、
前記請求項 2に記載の処理で、 LDGM (Low Density Generation Matrix)構造 の非正則なパリティ検査行歹 IJを生成することを特徴とする通信装置。
[12] 誤り訂正技術として、 LDPC (Low-Density Parity Check)符号を採用する通信シ ステムであって、
前記請求項 7に記載の処理で、所定の情報ビットを符号化する送信装置と、 既知の処理で符号語を復号する受信装置と、
を備えることを特徴とする通信システム。
[13] LDPC (Low-Density Parity Check)符号用のパリティ検査行列を用いて所定の 情報ビットを符号化する符号化器におレ、て、
前記請求項 7に記載の処理で、所定の情報ビットを符号化する符号化手段、 を備えることを特徴とする符号化器。
PCT/JP2007/065195 2006-08-04 2007-08-02 Procédé de génération de matrice d'inspection, procédé d'encodage, dispositif et système de communication et encodeur WO2008016117A1 (fr)

Priority Applications (4)

Application Number Priority Date Filing Date Title
JP2008527792A JP4901871B2 (ja) 2006-08-04 2007-08-02 検査行列生成方法、符号化方法、通信装置、通信システムおよび符号化器
EP07791869A EP2053751A4 (en) 2006-08-04 2007-08-02 INVESTIGATION MATRIX GENERATION PROCESS, CODING METHOD, COMMUNICATION DEVICE, COMMUNICATION SYSTEM AND CODIER
US12/376,344 US20100058140A1 (en) 2006-08-04 2007-08-02 Check-matrix generating method, encoding method, communication apparatus, communication system, and encoder
CA002660060A CA2660060A1 (en) 2006-08-04 2007-08-02 Check-matrix generating method, encoding method, communication apparatus, communication system, and encoder

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2006-213722 2006-08-04
JP2006213722 2006-08-04

Publications (1)

Publication Number Publication Date
WO2008016117A1 true WO2008016117A1 (fr) 2008-02-07

Family

ID=38997291

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2007/065195 WO2008016117A1 (fr) 2006-08-04 2007-08-02 Procédé de génération de matrice d'inspection, procédé d'encodage, dispositif et système de communication et encodeur

Country Status (7)

Country Link
US (1) US20100058140A1 (ja)
EP (1) EP2053751A4 (ja)
JP (1) JP4901871B2 (ja)
KR (1) KR20090040311A (ja)
CN (1) CN101502003A (ja)
CA (1) CA2660060A1 (ja)
WO (1) WO2008016117A1 (ja)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2010041466A1 (ja) * 2008-10-10 2010-04-15 パナソニック株式会社 符号化器、送信装置及び符号化方法
JP2012019561A (ja) * 2011-10-26 2012-01-26 Sony Corp 符号化方法および符号化装置
CN102957436A (zh) * 2011-08-17 2013-03-06 北京泰美世纪科技有限公司 一种低密度奇偶校验码译码装置和译码方法

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPWO2008132813A1 (ja) * 2007-04-17 2010-07-22 パナソニック株式会社 無線通信装置および無線通信方法
JP4858335B2 (ja) * 2007-07-10 2012-01-18 ソニー株式会社 符号化方法および符号化装置
CN102077471B (zh) * 2008-07-04 2014-03-12 三菱电机株式会社 校验矩阵生成装置、校验矩阵生成方法、校验矩阵生成程序、发送装置、接收装置以及通信系统
KR101125100B1 (ko) * 2010-12-03 2012-03-21 한국과학기술원 천공 기술을 활용하는 리드 솔로몬 기반 준순환 저밀도 패리티 검사 부호 생성 방법과 이를 이용하는 부호화 및 복호화 방법 및 저장 장치.
CN103034471B (zh) * 2012-12-10 2016-06-01 杜海洋 一种随机数的生成方法及系统
JP2016213701A (ja) * 2015-05-11 2016-12-15 富士通株式会社 誤り訂正方法、半導体装置、送受信モジュールおよび伝送装置
CN105429646B (zh) * 2015-06-30 2019-03-22 南京大学 一种咬尾阶梯码的编码及解码方法
WO2018012637A1 (ja) 2016-07-15 2018-01-18 シャープ株式会社 送信装置、受信装置、通信方法、および、集積回路
CN110380999B (zh) * 2018-04-12 2020-10-09 华为技术有限公司 概率非均匀调制的数据传输方法及装置
CN112136276B (zh) * 2018-05-29 2023-08-22 三菱电机株式会社 发送机、接收机、通信系统、编码率的变更方法、控制电路及存储介质
US11455208B2 (en) * 2020-08-20 2022-09-27 Western Digital Technologies, Inc. Soft information for punctured bit estimation in a data storage device
CN112800751B (zh) * 2021-02-19 2024-11-15 上海中通吉网络技术有限公司 移动端网页表单配置和校验方法、装置、设备和存储介质
US12301260B1 (en) * 2023-11-13 2025-05-13 Sk Hynix Nand Product Solutions Corp. Systems and methods for generating optimized combination sets for error correction in data transmission

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005110251A (ja) * 2003-09-26 2005-04-21 Mitsubishi Electric Research Laboratories Inc N個のサンプルを含む入力信号を、q元アルファベットから選択されたk個のシンボルのストリングに量子化する方法
WO2006027897A1 (ja) * 2004-09-03 2006-03-16 Mitsubishi Denki Kabushiki Kaisha Ldpc符号生成方法、通信装置および符号列生成方法

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080184084A1 (en) * 2004-10-08 2008-07-31 Mitsubishi Electric Corporation Check Matrix Generating Apparatus and Communication Apparatus
US7617439B2 (en) * 2005-01-10 2009-11-10 Broadcom Corporation Algebraic construction of LDPC (Low Density Parity Check) codes with corresponding parity check matrix having CSI (Cyclic Shifted Identity) sub-matrices
EP1965498B1 (en) * 2005-12-20 2014-06-25 Mitsubishi Electric Corporation Encoding of LDPC codes with an irregular parity check matrix obtained by masking
US20090063930A1 (en) * 2006-02-02 2009-03-05 Mitsubishi Electric Corporation Check matrix generating method, encoding method, decoding method, communication device, encoder, and decoder

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2005110251A (ja) * 2003-09-26 2005-04-21 Mitsubishi Electric Research Laboratories Inc N個のサンプルを含む入力信号を、q元アルファベットから選択されたk個のシンボルのストリングに量子化する方法
WO2006027897A1 (ja) * 2004-09-03 2006-03-16 Mitsubishi Denki Kabushiki Kaisha Ldpc符号生成方法、通信装置および符号列生成方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
M. FOSSORIER: "Quasi-Cyclic Low Density Parity Check Code", ISIT2003, 29 June 2003 (2003-06-29), pages 150
See also references of EP2053751A4 *

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9094044B2 (en) 2008-10-10 2015-07-28 Panasonic Intellectual Property Corporation Of America Encoder, transmitting apparatus, coding method and transmitting method
US9787325B2 (en) 2008-10-10 2017-10-10 Panasonic Intellectual Property Corporation Of America Encoder and coding method
US11848683B2 (en) 2008-10-10 2023-12-19 Panasonic Intellectual Property Corporation Of America Transmission apparatus
US11539378B2 (en) 2008-10-10 2022-12-27 Panasonic Intellectual Property Corporation Of America Integrated circuit for reception apparatus
US8473827B2 (en) 2008-10-10 2013-06-25 Panasonic Corporation Encoder, transmission device, and encoding method
US8607131B2 (en) 2008-10-10 2013-12-10 Panasonic Corporation Decoder, receiving apparatus, decoding method, and receiving method
JP2010114862A (ja) * 2008-10-10 2010-05-20 Panasonic Corp 符号化器、送信装置及び符号化方法
US8745473B2 (en) 2008-10-10 2014-06-03 Panasonic Corporation Encoder, transmitting apparatus, coding method and transmission method
WO2010041466A1 (ja) * 2008-10-10 2010-04-15 パナソニック株式会社 符号化器、送信装置及び符号化方法
US8954830B2 (en) 2008-10-10 2015-02-10 Panasonic Intellectual Property Corporation Of America Decoder, reception apparatus, decoding method, and reception method
US11043971B2 (en) 2008-10-10 2021-06-22 Panasonic Intellectual Property Corporation Of America Integrated circuit for transmission apparatus
US10224963B2 (en) 2008-10-10 2019-03-05 Panasonic Intellectual Property Corporation Of America Transmission apparatus, transmission method, reception apparatus, and reception method
US10819370B2 (en) 2008-10-10 2020-10-27 Panasonic Intellectual Property Corporation Of America Transmission apparatus, transmission method, reception apparatus and reception method
CN102957436B (zh) * 2011-08-17 2017-11-10 北京泰美世纪科技有限公司 一种低密度奇偶校验码译码装置和译码方法
CN102957436A (zh) * 2011-08-17 2013-03-06 北京泰美世纪科技有限公司 一种低密度奇偶校验码译码装置和译码方法
JP2012019561A (ja) * 2011-10-26 2012-01-26 Sony Corp 符号化方法および符号化装置

Also Published As

Publication number Publication date
US20100058140A1 (en) 2010-03-04
CA2660060A1 (en) 2008-02-07
JP4901871B2 (ja) 2012-03-21
KR20090040311A (ko) 2009-04-23
EP2053751A1 (en) 2009-04-29
EP2053751A4 (en) 2011-06-22
CN101502003A (zh) 2009-08-05
JPWO2008016117A1 (ja) 2009-12-24

Similar Documents

Publication Publication Date Title
WO2008016117A1 (fr) Procédé de génération de matrice d&#39;inspection, procédé d&#39;encodage, dispositif et système de communication et encodeur
US11616514B2 (en) Method and apparatus for channel encoding and decoding in a communication system using a low-density parity check code
JP4602418B2 (ja) 検査行列生成方法、符号化方法、復号方法、通信装置、符号化器および復号器
JP4598085B2 (ja) 検査行列生成方法
JP4620132B2 (ja) 検査行列生成方法、符号化方法、通信装置、通信システム、符号化器
EP1715588B1 (en) Algebraic construction of LDPC (Low Density Parity Check) codes with corresponding parity check matrix having CSI (Cyclic Shifted Identity) sub-matrices
CN1947368B (zh) 对具有可变块长度的块低密度奇偶校验码编码/解码的设备和方法
CN100568755C (zh) 用于编码和解码具有可变编码率的块低密度奇偶校验码的装置和方法
CN1993892B (zh) 用于编码和解码块低密度奇偶校验码的装置和方法
JP4987079B2 (ja) 検査行列生成装置
CN103152055B (zh) 编码和解码通信系统中的信道的设备和方法
US20090210767A1 (en) Apparatus and method for encoding and decoding channel in a communication system using low-density parity-check codes
CN102696175B (zh) 在使用线性分组码的通信系统中产生奇偶校验矩阵的装置和方法、发送/接收装置以及使用其的方法
KR20170075627A (ko) 통신 또는 방송 시스템에서 채널 부호화/복호화 방법 및 장치
WO2007092672A2 (en) Method and apparatus for encoding and decoding data
CN101252360A (zh) 一种高围长低码率多进制ldpc码的结构化构造方法
US8190977B2 (en) Decoder of error correction codes
KR20140145977A (ko) 패리티 검사 부호의 부호화 장치, 그의 부호화 방법, 복호화 장치 및 그의 복호화 방법
CN102811064B (zh) 一种多码率ldpc码的构造方法
KR20170060574A (ko) 통신 또는 방송 시스템에서 채널 부호화/복호화 방법 및 장치

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 200780029059.1

Country of ref document: CN

DPE2 Request for preliminary examination filed before expiration of 19th month from priority date (pct application filed from 20040101)
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 07791869

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2008527792

Country of ref document: JP

WWE Wipo information: entry into national phase

Ref document number: 2660060

Country of ref document: CA

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2007791869

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: RU

WWE Wipo information: entry into national phase

Ref document number: 1241/CHENP/2009

Country of ref document: IN

WWE Wipo information: entry into national phase

Ref document number: 12376344

Country of ref document: US

WWE Wipo information: entry into national phase

Ref document number: 1020097002239

Country of ref document: KR

DPE2 Request for preliminary examination filed before expiration of 19th month from priority date (pct application filed from 20040101)
点击 这是indexloc提供的php浏览器服务,不要输入任何密码和下载