CN101567697B - Coder and method for coding rate-compatible low-density parity-check codes - Google Patents
Coder and method for coding rate-compatible low-density parity-check codes Download PDFInfo
- Publication number
- CN101567697B CN101567697B CN 200910085516 CN200910085516A CN101567697B CN 101567697 B CN101567697 B CN 101567697B CN 200910085516 CN200910085516 CN 200910085516 CN 200910085516 A CN200910085516 A CN 200910085516A CN 101567697 B CN101567697 B CN 101567697B
- Authority
- CN
- China
- Prior art keywords
- matrix
- submatrix
- diagonal
- basic extension
- row
- Prior art date
- Legal status (The legal status 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 status listed.)
- Expired - Fee Related
Links
Images
Abstract
The invention discloses a method for obtaining parity check matrix of rate-compatible low-density parity-check codes, comprising the following steps of: selecting a mother code matrix and determining the size of a basic extension matrix; dividing the basic extension matrix into six sub-matrixes (see the formula at right side); according to a minimum ring maximum principle, respectively determiningthe value of each sub-matrix; subsequently combining each sub-matrix to form the basic extension matrix; finally correcting the obtained basic extension matrix by using a current extension factor z and extending the corrected basic extension matrix by using a matrix P with the size of z*z, thus obtaining a parity check matrix; and using the parity check matrix to carry out the LDPC coding to thedata to be coded. When the method and the coder are applied,on one hand, the rate compatibility can be realized; on the other hand, as the minimum and maximum principle is adopted to determine the value of each sub-matrix in each basic extension matrix, the decoding performance can be improved compared to the LDPC codes realizing rate-compatibility by a punching way.
Description
Technical field
The present invention relates to the coding techniques of low density parity check code (LDPC), particularly a kind of LDPC coding method and encoder of rate-compatible.
Background technology
(Low-Density Parity-Check Codes is one type of linear block codes LDPC) to low density parity check code, and its check matrix can be represented with binary matrix; As shown in Figure 1, each shows j 1 (being that the column weight amount is j), and each row has k 1 (weight be k at once); And other elements all are zero; J >=3 wherein, k>j, j, k are less integer.The check matrix of LDPC can also be represented with a kind of two-dimensional plot, is commonly referred to as Tanner figure now.Tanner figure is directly corresponding with the binary system check matrix.As shown in Figure 2, be variable node (s below among the Tanner figure
0, s
1..., s
9), can think bit or the row in the check matrix in the code word; Be check-node (c above
0, c
1..., c
4), each node is represented the delegation in a check equations or the check matrix.When a certain bit in the code word is included in a certain check equations, there is line among Fig. 2 between the respective nodes.For each node, the number on the limit that is attached thereto is called the degree of this node (degree).The structure of low-density check matrix has conclusive effect for the performance of sign indicating number.Different building methods all is in order to realize following purpose: optimize the non-regular code node number of degrees and distribute; Increase the cycle among the figure, reduce encoder complexity.If the node among the Tanner figure is built into form as shown in Figure 3, then the becate cycle among the figure is 4.
At present, quasi-cyclic LDPC sign indicating number is applied in the practical communication field more and more.The check matrix of this yard is that a size is (m
b* z) * (n
b* z) matrix representes that with H wherein, z representes spreading factor.This check matrix H is that the cycle specificity matrix of z * z constitutes with complete zero battle array by many sizes.The generation type of concrete this check matrix is: definition P is a shift matrix that size is the unit matrix of z * z, and its form is following:
The cyclic shift that above-mentioned P matrix carries out different side-play amounts obtains the cycle specificity matrix that a plurality of sizes are z * z, and check matrix H just is made up of these cycle specificity matrixes and complete zero battle array.The form of concrete check matrix is shown below:
Wherein,
Expression is carried out h to matrix P
Ij bThe matrix that obtains after the inferior cyclic shift, if
Then
Definition
Then check matrix can be expressed as
, wherein, the size of check matrix H is M * N, and N=n
b* z, M=m
b* z.With H
bThe basis matrix that is called check matrix just can obtain check matrix according to this basis matrix with replacement matrix P.
Basis matrix h
bConstitute by two parts,
H wherein
B1Corresponding to system bits bit, H
B2Corresponding to the check digit bit.
(Rate-Compatible, LDPC sign indicating number RC) has important role to rate-compatible in the block data communication of reality.A major advantage of rate-compatible code is exactly only to need a pair of volume, decoder can accomplish coding, the work decoding of each speed.Except advantage with low complex degree, rate-compatible code also be automatic re-send request may (Automatic Repeat Request, ARQ)/(ForwardError Correction, FEC) agreement provides an effective framework in forward error correction.
Present rate-compatible code all is that this mode can not guarantee the performance of the LDPC sign indicating number in a very big range of code rates through the coding and decoding operation of the different rates of hole knockout realization.This is because under the situation of high code check, has destroyed the soft information that most of decoder needs, thereby has caused decoding failure.
Summary of the invention
In view of this, the present invention provides a kind of LDPC coding method and encoder of rate-compatible, utilizes the LDPC coding that is carried out rate-compatible by this method and encoder, the coding efficiency that can greatly improve.
For realizing above-mentioned purpose, the present invention adopts following technical scheme:
A kind of check matrix acquisition methods of low density parity check code of rate-compatible comprises:
Select female sign indicating number matrix, and the size (m of definite basic extension matrix
b+ Δ m) * (n
b+ Δ m), wherein, Δ m=Δ xm
b,
Expression rounds up, R
0Be female sign indicating number matrix corresponding code rate, m
bAnd n
bBe respectively the line number and the columns of female sign indicating number matrix, R is a target bit rate;
With basic extension matrix H
ExtensionBe divided into six sub-matrices,
Wherein, H
b SystemBe the information submatrix of said female sign indicating number matrix, H
b ParityBe the syndrome matrix of said female sign indicating number matrix, H
Upper_right, H
Lower_left, H
Lower_mid, H
Lower_rightSize is respectively m
b* Δ m, Δ m * (n
b-m
b), Δ m * m
b, Δ m * Δ m;
With submatrix H
Upper_rightAll elements be set to-1;
For submatrix H
Lower_mid, all elements on the preset diagonal is set to 0, and all the other elements are set to-1, and said preset diagonal is for connecting the
Row the 0th column element and nm
b-1 row
The diagonal of column element, and connect (n-1) m
bRow
Element and
Row m
bThe diagonal of-1 column element, wherein, n=1 ..., Δ x, and n is a positive integer;
The constructor matrix H
Lower_rightMask code matrix H
MASK Lower_rightFor possessing many cornerwise lower triangular matrixs, comprise mask code matrix H in said many diagonal
MASK Lower_rightLeading diagonal and Δ x bar counter-diagonal, wherein, said Δ x bar counter-diagonal is said matrix H
Lower_midDescribed in preset cornerwise extended line, all elements on said many diagonal is set to 1, all the other elements are set to 0;
According to submatrix H
b System, H
b Parity, H
Upper_right, H
Lower_midWith submatrix H
Lower_rightMask code matrix confirm submatrix H
Lower_leftMask code matrix H
MASK Lower_leftDegree distribute structure mask code matrix H
MASK Lower_left
With mask code matrix H
MASK Lower_rightIn all values be that 0 element is at submatrix H
Lower_rightMiddle corresponding element is set to " 1 ", with matrix H
MASK Lower_rightElement on the middle leading diagonal is at submatrix H
Lower_rightMiddle corresponding element is set to " 0 ", with mask code matrix H
MASK Lower_leftIn all values be that 0 element is at submatrix H
Lower_leftMiddle corresponding element is set to " 1 "; According to the long maximization of minimum ring criterion, be matrix H
MASK Lower_rightElement on the middle counter-diagonal is in matrix H
Lower_rightMiddle corresponding element and matrix H
MASK Lower_leftThe element that middle value is " 1 " is at submatrix H
Lower_leftMiddle corresponding element is provided with corresponding side-play amount;
Six sub-matrices of confirming are merged obtain said basic extension matrix;
Utilize current spreading factor z that said basic extension matrix is carried out correcting process, and utilize big or small replacement matrix P that revised basic extension matrix is expanded, obtain check matrix for z * z, wherein,
Expression rounds up, and K is the information bit length of present encoding, k
b=n
b-m
b
The check matrix that utilization obtains carries out the LDPC coding to data to be encoded.
Preferably, minimize criterion, structure mask code matrix H according to long maximization of minimum ring and becate number
MASK Lower_left, satisfy minimum ring and grow up in equaling 6.
Preferably, according to
With
Configuration integer quotient C
1..., C
9, D
1..., D
9, the minimum ring of said basic extension matrix is grown up in equaling 6, wherein, C
1..., C
9Be nonnegative integer, z
MaxBe the maximum of spreading factor, h
I ' j ' Lower_rightBe submatrix H
Lower_rightIn the element value of the i ' row j ' row, and i ' and j ' be the index of element on the said counter-diagonal, h
Ij Lower_leftBe submatrix H
Lower_leftIn the element value of the capable j of i row, and i and j are submatrix H
Lower_leftIn with mask code matrix H
MASK Lower_leftThe corresponding element index of element that middle value is " 1 ".
Preferably, said correcting process is:
Wherein, h
Ij bBe basic extension matrix H
ExtensionIn the element value of the capable j of i row, (h
Ij b)
ModifiedBe h
Ij bRevised element value.
Preferably, in advance according to a plurality of optional target bit rate select target matrixes and the basic extension matrix under definite each target bit rate;
When carrying out the LDPC coding, according to the corresponding basic extension matrix of current target code rate selection, and carry out said correcting process and extended operation, obtain check matrix.
A kind of encoder of low density parity check code of rate-compatible comprises:
Basic extension matrix is confirmed the unit, is used to select female sign indicating number matrix, and the size (m of definite basic extension matrix
b+ Δ m) * (n
b+ Δ m), wherein, Δ m=Δ xm
b,
Expression rounds up, R
0Be female sign indicating number matrix corresponding code rate, m
bAnd n
bBe respectively the line number and the columns of female sign indicating number matrix, R is a target bit rate; With basic extension matrix H
ExtensionBe divided into six sub-matrices,
Wherein, H
b SystemBe the information submatrix of said female sign indicating number matrix, H
b ParityBe the syndrome matrix of said female sign indicating number matrix, H
Upper_right, H
Lower_left, H
Lower_mid, H
Lower_rightSize is respectively m
b* Δ m, Δ m * (n
b-m
b), Δ m * m
b, Δ m * Δ m; With submatrix H
Upper_rightAll elements be set to-1; For submatrix H
Lower_mid, all elements on the preset diagonal is set to 0, and all the other elements are set to-1, and said preset diagonal is for connecting the
Row the 0th column element and nm
b-1 row
The diagonal of column element, and connect (n-1) m
bRow
Element and
Row m
bThe diagonal of-1 column element, wherein, n=1 ..., Δ x, and n is a positive integer; The constructor matrix H
Lower_rightMask code matrix H
MASK Lower_rightFor possessing many cornerwise lower triangular matrixs, comprise mask code matrix H in said many diagonal
MASK Lower_rightLeading diagonal and Δ x bar counter-diagonal, wherein, said Δ x bar counter-diagonal is said matrix H
Lower_midDescribed in preset cornerwise extended line, all elements on said many diagonal is set to 1, all the other elements are set to 0; According to submatrix H
b System, H
b Parity, H
Upper_right, H
Lower_midWith submatrix H
Lower_rightMask code matrix confirm submatrix H
Lower_leftMask code matrix H
MASK Lower_leftDegree distribute structure mask code matrix H
MASK Lower_leftWith mask code matrix H
MASK Lower_rightIn all values be that 0 element is at submatrix H
Lower_rightMiddle corresponding element is set to " 1 ", with matrix H
MASK Lower_rightElement on the middle leading diagonal is at submatrix H
Lower_rightMiddle corresponding element is set to " 0 ", with mask code matrix H
MASK Lower_leftIn all values be that 0 element is at submatrix H
Lower_leftMiddle corresponding element is set to " 1 "; According to the long maximization of minimum ring criterion, be matrix H
MASK Lower_rightElement on the middle counter-diagonal is in matrix H
Lower_rightMiddle corresponding element and matrix H
MASK Lower_leftThe element that middle value is " 1 " is at submatrix H
Lower_leftMiddle corresponding element is provided with corresponding side-play amount; Six sub-matrices of confirming are merged obtain said basic extension matrix;
The correcting process unit is used to utilize current spreading factor z that said basic extension matrix is confirmed that the basic extension matrix that the unit obtains carries out correcting process;
Expanding element is used for utilizing big or small replacement matrix P for z * z that the revised basic extension matrix that said correcting process unit obtains is expanded, and obtains check matrix, wherein,
Expression rounds up, and K is the information bit length of present encoding, k
b=n
b-m
b
Coding unit, the check matrix that is used to utilize said expanding element to obtain carries out the LDPC coding to data to be encoded.
Preferably, said basic extension matrix confirms that the unit comprises definite subelement and storing sub-units,
Said definite subelement; Be further used for based on a plurality of optional target bit rate select target matrixes and the basic extension matrix under definite each target bit rate, and the target bit rate that each is optional is kept in the said storing sub-units with basic extension matrix is corresponding accordingly;
Said correcting process unit is further used in said storing sub-units, shifting to an earlier date the basic extension matrix corresponding with the current target code check.
Visible by technique scheme, among the present invention, select female sign indicating number matrix, and the size of definite basic extension matrix; Basic extension matrix is divided into six sub-matrices,
According to the minimum ring maximization principle, confirm the value of each sub-matrices respectively, again each sub-matrices is merged the formation base extended matrix; At last, utilize current spreading factor z that the basic extension matrix that obtains is revised, utilize big or small matrix P that revised basic extension matrix is expanded again, obtain check matrix for z * z; Utilize this check matrix that data to be encoded are carried out the LDPC coding.Through the LDPC sign indicating number that aforesaid way obtains, on the one hand, can realize rate-compatible; On the other hand, minimize the value that maximization principle is confirmed each sub-matrices in each basic extension matrix owing to adopt, therefore; LDPC sign indicating number with respect to hole knockout realization rate-compatible can improve decoding performance.
Description of drawings
Fig. 1 utilizes binary matrix to represent the sketch map of LDPC code check matrix.
Fig. 2 is the sketch map that utilizes Tanner figure expression LDPC code check matrix.
Fig. 3 is the flow chart of the check matrix acquisition methods of LDPC among the present invention.
Fig. 4 is the structural representation one of basic extension matrix among the present invention.
Fig. 5 is the structural representation two of basic check matrix among the present invention.
Fig. 6 is the FER (BLER) of the LDPC sign indicating number of constructing in the present invention and the prior art and the performance comparison curved line relation between the signal to noise ratio (Eb/No).
Fig. 7 is LDPC sign indicating number and the FER (FER) of existing Turbo code and the performance comparison curved line relation between the signal to noise ratio (Eb/No) among the present invention.
Embodiment
For making the object of the invention, technological means and advantage clearer, the present invention is explained further details below in conjunction with accompanying drawing.
Basic thought of the present invention is: according to target bit rate; According to the long maximization of minimum ring criterion female sign indicating number matrix is expanded to the basic extension matrix that satisfies the requirement of different target code check; To realize the LDPC coding of rate-compatible, reduce thereby avoid handling the decoding performance that causes owing to punching.
The LDPC encoder of rate-compatible is provided among the present invention, has comprised: basic extension matrix is confirmed unit, correcting process unit, expanding element and coding unit.
In above-mentioned LDPC encoder, LDPC coding method that can embodiment of the present invention.Fig. 3 is the flow chart of the coding method of LDPC among the present invention.As shown in Figure 3, this method comprises:
Step 301 based on the requirement of code check R and code length N, is selected female sign indicating number matrix, and the size of definite basic extension matrix.
This step can realize through following step:
(11), confirm information bit length K and check bit length M according to the requirement of code check R and code length N; Wherein, information bit length K=RN, check bit length M=N-K, in the formula, information bit length K, code length N and check bit M are positive integer;
Suppose that in the present invention's one specific embodiment, target bit rate is 1/3, code length N has three kinds of requirements, is respectively 3180,6060,15342, then corresponding information bit length K=1060,2020,5114, corresponding check bit length M=2120,4040,10228; Below be that example specifies each step just with this kind situation.
(12) selecting size is m
b* n
bFemale sign indicating number matrix H
b Uniform, concrete female sign indicating number matrix selection mode is identical with existing implementation.Wherein, female sign indicating number matrix is m by size
b* (n
b-m
b) the systematic bit subpacket matrix H
b SystemWith size be m
b* m
bCheck digit submatrix H
b ParityTwo parts are formed, promptly
H wherein
b ParityIt is the accurate lower triangular matrix of double diagonal line;
In specific embodiment, female sign indicating number matrix line number of selection is m
b=16, columns n
b=32, female sign indicating number code check is 1/2, and female sign indicating number matrix is 16 * 16 systematic bit subpacket matrix H by size
b SystemWith size be 16 * 16 check digit submatrix H
b SystemForm, wherein H
b ParityIt is the accurate lower triangular matrix of double diagonal line;
Wherein, systematic bit subpacket matrix
The check digit submatrix
(13), need to confirm the basic extension matrix H of structure according to code check R
ExtensionLine number and columns, this basic extension matrix H
ExtensionLine number is m
b+ Δ m, columns are n
b+ Δ m, Δ m=m
bΔ x;
Wherein, by R=((n
b+ m
bΔ x)-(m
b+ m
bΔ x))/(n
b+ m
bAnd R Δ x),
0=(n
b-m
b)/n
b, can draw
Expression rounds up, R
0Be female sign indicating number matrix corresponding code rate.Can be R like this, just according to code check
0A female sign indicating number matrix size expand the size that code check is the basic extension matrix of R.Certainly, this step has only been confirmed the size of basic extension matrix, also needs the further concrete value of confirming basic extension matrix at subsequent step, just can obtain the check matrix that code check is R through this basic extension matrix is replaced then;
In specific embodiment, because the target bit rate of LDPC is R=1/3, then the code check of basic extension matrix also is R=1/3, again according to female sign indicating number code check R
0=1/2, can get Δ x=1, Δ m=m
b=16, thereby draw basic extension matrix H
ExtensionLine number be 32, columns is 48.
Step 302 is with basic extension matrix H
ExtensionBe divided into six sub-matrices
Female sign indicating number matrix expansion by selecting obtains basic extension matrix; Therefore; With the matrix of female sign indicating number matrix as the upper left corner in the basic extension matrix; And in view of the above basic extension matrix is divided into a plurality of submatrixs, confirm the value of other submatrixs except that female sign indicating number matrix more one by one, thereby finally confirm basic extension matrix.
Particularly, with basic extension matrix H
ExtensionBe divided into six sub-matrices, promptly
Wherein, H
b SystemBe the information submatrix of female sign indicating number matrix, H
b ParityBe the syndrome matrix of female sign indicating number matrix, H
Upper_right, H
Lower_left, H
Lower_mid, H
Lower_rightBe the submatrix that expands, size is respectively m
b* Δ m, Δ m * (n
b-m
b), Δ m * m
b, Δ m * Δ m.
In specific embodiment, basic extension matrix H
ExtensionAs shown in Figure 4, matrix comprises 5 parts, and wherein, the upper left corner is that size is female sign indicating number matrix H of 16 * 32
b Uniform,
Expansion H
Upper_right, H
Lower_left, H
Lower_mid, H
Lower_rightMatrix size is 16 * 16 respectively.
Step 303 is with submatrix H
Upper_rightIn all elements be set to-1.
Step 304 is for submatrix H
Lower_mid, all elements on the preset diagonal is set to 0, and all the other elements are set to-1.
This step is confirmed submatrix H
Lower_midThe value of middle all elements.As shown in Figure 5, submatrix H
Lower_midThe form that presents double diagonal line.All elements on each diagonal is set to 0, and all the other elements are set to-1.Concrete diagonal is: connect the
Row the 0th column element and nm
b-1 row
The diagonal of column element and, connect (n-1) m
bRow
Element and
Row m
bThe diagonal of-1 column element.Wherein, the value of n is all positive integers smaller or equal to Δ x, and promptly n gets since 1, gets Δ x always, when n gets arbitrary value, obtains two diagonal, so submatrix H
Lower_midHave 2 Δ x bar diagonal.Like this, promptly can confirm submatrix H
Lower_midThe all elements value, the final submatrix H that confirms
Lower_midRow heavily is 1, and column weight is Δ x.
In specific embodiment, size is 16 * 16 matrix H in the middle of structure foundation extended matrix the latter half
Lower_midThe time, at first all elements in the matrix is changed to " 1 ", again the element on the diagonal is changed to " 0 ", here because Δ x=1, so n only gets 1, thus article one diagonal connects is submatrix H
Lower_midThe element of the element of middle eighth row the 0th row and the 15th row the 7th row, that the second diagonal connects is submatrix H
Lower_midIn the element that is listed as of element and the 7th row the 15th of the 0th row the 8th row; The row of this submatrix heavily is 1, and column weight also is 1;
Step 305 is according to submatrix H
Lower_midConstruct a submatrix H with the basic extension matrix lower right corner
Lower_rightThe mask code matrix H corresponding and structure is identical
MASK Lower_right
The mask code matrix H of this step structure
MASK Lower_rightIn element value be " 0 " or " 1 ", wherein, " 0 " expression does not have a matrix operation, " 1 " expression has matrix operation.
Particularly, submatrix H
Lower_rightMask code matrix H
MASK Lower_rightFor possessing many cornerwise lower triangular matrixs, wherein, comprise mask code matrix H in many diagonal
MASK Lower_rightLeading diagonal and Δ x bar counter-diagonal, wherein, Δ x bar counter-diagonal is submatrix H
Lower_midIn all cornerwise extended lines, as shown in Figure 5.And, at mask code matrix H
MASK Lower_rightIn, all elements on each diagonal is set to 1, and all the other elements are set to 0.
In specific embodiment, mask code matrix is a lower triangular matrix with double diagonal line, and counter-diagonal is by submatrix H
Lower_midArticle one diagonal prolong and to obtain, concrete this mask code matrix
Step 306 is according to submatrix H
b System, H
b Parity, H
Upper_right, H
Lower_midWith submatrix H
Lower_rightMask code matrix confirm submatrix H
Lower_leftMask code matrix H
MASK Lower_leftDegree distribute structure mask code matrix H
MASK Lower_left
In this step, can adopt the density evolution rationale structure of definite several sub-matrices, confirm submatrix H
Lower_leftMask code matrix H
MASK Lower_leftDegree distribute, adopt progressive edged method construct mask code matrix H
MASK Lower_leftThis mask code matrix H
MASK Lower_leftWith mask code matrix H
MASK Lower_rightSimilar, element value wherein is " 0 " or " 1 ", and " 0 " expression does not have matrix operation, and " 1 " expression has matrix operation.
Preferably, at structure mask code matrix H
MASK Lower_leftProcess in, carry out according to long maximization of minimum ring and the minimized criterion of becate number.
In specific embodiment, according to the structure of 4 parts of having constructed, adopt density evolution theoretical, obtain lower left corner mask code matrix H
MASK Lower_leftColumn weight distribution be 0.6875x
0+ 0.125x
2+ 0.1875x
16, just the column weight amount is respectively 0,2,16 columns and accounts for total columns ratio and be respectively 0.6875,0.125,0.1875; Adopt progressive bordering method edged to make basic extension matrix H again
ExtensionThe long maximization of minimum ring, promptly minimum ring is long by>=6, and the becate number minimizes, the mask code matrix that obtains does
Step 307 according to the long maximization of minimum ring criterion, is matrix H
MASK Lower_rightElement on the middle counter-diagonal is in matrix H
Lower_rightMiddle corresponding element and matrix H
MASK Lower_leftThe element that middle value is " 1 " is at submatrix H
Lower_leftMiddle corresponding element is provided with corresponding side-play amount, and with matrix H
Lower_rightAnd H
Lower_leftIn all the other elements be set to " 1 ".
In this step, according to mask code matrix H
MASK Lower_rightAnd H
MASK Lower_left, confirm corresponding submatrix H
Lower_rightAnd H
MASK Lower_leftThe element value.Particularly, according to
With
Configuration integer quotient C
1..., C
9, D
1..., D
9, the minimum ring of basic extension matrix is grown up in equaling 6, wherein, C
1..., C
9Be nonnegative integer, z
MaxBe the maximum of spreading factor, h
I ' j ' Lower_rightBe submatrix H
Lower_rightIn the element value of the i ' row j ' row, and i ' and j ' are submatrix H
Lower_rightThe index of element on the middle counter-diagonal, h
Ij Lower_leftBe submatrix H
Lower_leftIn the element value of the capable j of i row, and i and j are submatrix H
Lower_leftIn with mask code matrix H
MASK Lower_leftThe corresponding element index of element that middle value is " 1 ".With matrix H
Lower_rightAnd H
Lower_leftIn all the other elements be set to " 1 ".
In specific embodiment, at first dispose lower left corner submatrix H
Lower_leftSide-play amount: at H
Lower_leftIn, will be corresponding to mask code matrix H
MASK Lower_leftMiddle element is all filled " 1 " for the position of " 0 ", and remainder adopts following computing formula to obtain:
Next dispose lower right corner submatrix H
Lower_rightSide-play amount: at H
Lower_rightIn, earlier the element on the leading diagonal is " 0 " for filling out, adopt following formula to calculate the element on the counter-diagonal again:
In the formula, h
I, j Lower_rightBe expansion lower right corner matrix sub H
Lower_leftIn the element of capable, the j of i row, and i=8,9 ..., 15, j=i-8; This element satisfies on counter-diagonal; z
MaxBe the obtainable maximum of spreading factor z, z
Max=384, mod is the complementation oeprator, at last all the other position elements is changed to " 1 ";
Should
Step 308 is merged into basic extension matrix with six sub-matrices of confirming.
So far, promptly can obtain basic extension matrix.The operation of above-mentioned steps 301~308 can be accomplished corresponding function by the basic extension matrix unit in the aforementioned encoder.
Step 309, according to current spreading factor z to basic extension matrix H
ExtensionCarry out correcting process, obtain revised basic extension matrix H
Modified
Female sign indicating number matrix is the spreading factor z according to maximum
MaxObtain, and in the above-mentioned process of confirming each sub-matrices, be based on also that female sign indicating number matrix and the largest extension factor carry out, so the basic extension matrix that merging obtains in the step 308 also is corresponding z with the largest extension factor
Max.And in practical application, the spreading factor z that encodes maybe be less than largest extension factor z
Max, therefore, need revise basic extension matrix, so that this revised basic extension matrix H according to true extension factor z
ModifiedIn the value of all elements all less than the spreading factor z of LDPC sign indicating number.
Preferably, concrete correcting mode is:
Wherein, n
Ij bBe basic extension matrix H
ExtensionIn the element value of the capable j of i row, (h
Ij b)
ModifiedBe h
Ij bRevised element value.
In specific embodiment, information bit length is respectively 1060,2020,5114, corresponding spreading factor
Be respectively 67,127,320, thereby obtain 3 revised matrix H
1060 Modified, H
2020 Modified, H
5114 ModifiedShould
Should
Should
The operation of this step can be accomplished corresponding function by the correcting process unit in the aforementioned LDPC encoder.
Step 310 is utilized the replacement matrix P of size for z * z, revised basic extension matrix is expanded required (m when obtaining actual coding
b+ Δ m) * capable (n of z
b+ Δ m) * check matrix H of z row.
The mode of specifically expanding just repeats no more with existing identical here.
In specific embodiment, information bit length is respectively 1060,2020,5114, corresponding spreading factor
Be respectively 67,127,320, be respectively 2144 * 3216,4064 * 6096,10240 * 15360 check matrix H thereby obtain 3 sizes
1060, H
2020, H
5114
The operation of this step can be accomplished corresponding function by the expanding element in the aforementioned LDPC encoder.
Step 311, the check matrix H of utilizing step 310 to confirm is carried out the LDPC coding to data to be encoded.
Next, just can utilize the check matrix that obtains to carry out the LDPC coding.Following brief account is cataloged procedure once, and still the parameter with above-mentioned specific embodiment is an example:
Code check R=1/3 is divided into N-2M information bit s, a M check bit c to a code word x
1With M check bit c
2, x=[s, c
1, c
2], correspondingly, be divided into 6 sparse matrixes to the check matrix H of 2M * N,
Be H
1, H
2, H
3, H
4, H
5And H
6, size is respectively M * (N-2M), M * M, M * M, M * (N-2M), M * M, M * M, according to Hx
T=0
T, can get:
So can obtain
Because H
3Be full null matrix, release
Again by
Further release
Because matrix H
2Adopt the accurate lower triangular structure of double diagonal line, H
6Adopt strict lower triangular structure, so H
2 -1And H
6 -1Have very simple form, can directly obtain check bit part c according to the top computing formula
1And c
2, and can guarantee that encoder has linear complexity.
The concrete operations step of utilizing check matrix H to encode adopts existing mode, and is specific as follows:
Step 1: with k
x=k
bBefore * z-K 0 bit is added in the block of information of K bit, constitute k
b* z bit information bit s; For example, when K=1060, k
x=k
b* z-K=16 * 67-1060=12;
Step 2: to the submatrix [H in the check matrix H
1| H
2] carry out the elementary row conversion, with check matrix [H
1| H
2] in all provisional capitals be added in first row, calculate first of check code;
Because submatrix H
2Be lower triangular matrix, therefore at first count, as above calculate since first row.
Step 3: according to submatrix [H
1| H
2] calculating other check digit with recurrence method, the check code that calculates with step 2 constitutes c
1
Step 4: according to submatrix [H
4| H
5| H
6] calculate check digit c with the method for recursion
2
Step 5: 0 bit that deletion is added in step 1 in s.If the remaining bit number of code word x is not equal to N, then the y=2m of code word x afterbody
bThe individual check bit of * z-N+K (y is a less number) needs deletion, for example, when K=1060, y=2m
b* z-N+K=2 * 16 * 67-3180+1060=24.
Like this, promptly accomplished and utilized the check matrix that obtains according to the inventive method to carry out the process of LDPC coding.The above-mentioned operation that utilizes check matrix to carry out the LDPC coding can be accomplished corresponding function by the coding unit in the aforementioned LDPC encoder.
So far, the LDPC coding method flow process of rate-compatible finishes among the present invention.By the way, can obtain the check matrix under the various code checks according to suitable female sign indicating number matrix, thereby realize the LDPC coding of rate-compatible.
In the concrete implementation procedure of the invention described above, can be directed against various optional target bit rates, execution in step 301~308 is then with basic extension matrix that obtains and the corresponding corresponding preservation of target bit rate.When carrying out actual LDPC coding; Require the select target code check according to current code check, extract the basic extension matrix corresponding with this target bit rate again, execution in step 309~310 is carried out correcting process and expansion; Obtain corresponding check matrix, utilize step 311 to carry out the LDPC coding again.Like this, the operation of abovementioned steps 301~308 can be carried out and preserve in advance, and the real-time operation amount in the time of can reducing the LDPC coding improves coding rate.
Correspondingly; When adopting the above-mentioned mode of confirming corresponding basic extension matrix to various optional target bit rates in advance; Concrete realization in encoder can be confirmed that the unit comprises for: the basic extension matrix in the encoder and confirm subelement and storing sub-units, wherein, confirm that subelement is used for to various optional target bit rates; Execution in step 301~308 is kept at the basic extension matrix that obtains in the storing sub-units with target bit rate is corresponding accordingly then.The basic extension matrix corresponding with the current goal code check preferably, confirmed to extract in the memory cell the unit from basic extension matrix in the correcting process unit when obtaining the basic extension matrix that basic extension matrix confirms that the unit obtains.
Concrete implementation procedure by the invention described above is visible, and in the LDPC sign indicating number process of structure rate-compatible, employing is expanded by a female sign indicating number matrix and obtained basic extension matrix among the present invention, expands the mode that obtains check matrix by basic extension matrix again.Aforesaid way is compared with existing hole knockout, has kept the useful information of decoding as far as possible, and in the process of confirming basic extension matrix, carries out according to the long maximized criterion of minimum ring, therefore can improve decoding performance.
In order to identify the transmission performance of the LDPC sign indicating number that forms according to mode of the present invention; Adopt the transmission performance of the LDPC sign indicating number that the awgn channel model constructs the present invention to carry out l-G simulation test, and with prior art in the LDPC sign indicating number realized through hole knockout carried out performance relatively.Table 1 is for being used to test the situation that is provided with each parameter of embodiment of analytical performance.
The parameter title | Numerical value |
Information bit long (K) | 1024 |
|
2/3 |
Modulating mode | BPSK |
Channel model | AWGN |
Decoded mode | LDPC:Log-BP maximum iteration time=50 |
Fig. 6 is the FER (BLER) of the LDPC sign indicating number of constructing in the present invention and the prior art and the performance comparison curved line relation between the signal to noise ratio (Eb/No).In Fig. 6, dotted line is that solid line is according to the transmission performance curve of the LDPC sign indicating number of method realization of the present invention according to the transmission performance curve of the LDPC sign indicating number of hole knockout realization in the prior art.Visible by Fig. 6, with existing punching compared with techniques, adopt the FER performance of the LDPC sign indicating number of the present invention's structure to be superior to prior art.
Simultaneously, also the LDPC sign indicating number of constructing according to mode of the present invention and the transmission performance of existing Turbo code emulation and comparison have been carried out.Table 2 is for being used to test the situation that is provided with each parameter of embodiment of analytical performance.
The parameter title | Numerical value |
Information bit long (K) | 1060、2020、5114 |
|
1/3 |
Modulating mode | BPSK |
Channel model | AWGN |
Decoded mode | Turbo:Log-MAP maximum iteration time=8 LDPC:Log-BP maximum iteration time=50 |
Fig. 7 is LDPC sign indicating number and the FER (FER) of existing Turbo code and the performance comparison curved line relation between the signal to noise ratio (Eb/No) among the present invention.In Fig. 7, dotted line is the transmission performance curve of Turbo code, and solid line is the transmission performance curve of LDPC sign indicating number among the present invention.Visible by Fig. 7, the FER of the LDPC sign indicating number of employing the present invention structure is functional: in information bit length is 2020, and FER is 10
-3The time, the LDPC sign indicating number manys the gain of 0.08dB than Turbo code; In information bit length is 5114 o'clock, and FER is 10
-3The time, the LDPC sign indicating number manys the gain of 0.1dB than Turbo code; And at FER is 10
-4The time, the LDPC sign indicating number can overcome the floor effect that Turbo code exists.
More than being merely preferred embodiment of the present invention, is not to be used to limit protection scope of the present invention.All within spirit of the present invention and principle, any modification of being done, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (7)
1. the check matrix acquisition methods of the low density parity check code of a rate-compatible is characterized in that, this method comprises:
Select female sign indicating number matrix, and the size (m of definite basic extension matrix
b+ Δ m) * (n
b+ Δ m), wherein, Δ m=Δ xm
b,
Expression rounds up, R
0Be female sign indicating number matrix corresponding code rate, m
bAnd n
bBe respectively the line number and the columns of female sign indicating number matrix, R is a target bit rate;
With basic extension matrix H
ExtensionBe divided into six sub-matrices,
Wherein,
Be the information submatrix of said female sign indicating number matrix,
Be the syndrome matrix of said female sign indicating number matrix, H
Upper_right, H
Lower_left, H
Lower_mid, H
Lower_rightSize is respectively m
b* Δ m, Δ m * (n
b-m
b), Δ m * m
b, Δ m * Δ m;
With submatrix H
Upper_rightAll elements be set to " 1 ";
For submatrix H
Lower_mid, all elements on the preset diagonal is set to " 0 ", and all the other elements are set to " 1 ", and said preset diagonal is for connecting the
Row the 0th column element and nm
b-1 row
The diagonal of column element, and connect (n-1) m
bRow
The element of row and the
The diagonal of row mb-1 column element, wherein, n=1 ..., Δ x, and n is a positive integer;
The constructor matrix H
Lower_rightMask code matrix
For possessing many cornerwise lower triangular matrixs, comprise mask code matrix in said many diagonal
Leading diagonal and Δ x bar counter-diagonal, wherein, said Δ x bar counter-diagonal is said matrix H
Lower_midDescribed in preset cornerwise extended line, all elements on said many diagonal is set to " 1 ", all the other elements are set to " 0 ";
According to submatrix
H
Upper_right, H
Lower_midWith submatrix H
Lower_rightMask code matrix confirm submatrix H
Lower_leftMask code matrix
Degree distribute the structure mask code matrix
With mask code matrix
In all values be that the element of " 0 " is at submatrix H
Lower_rightMiddle corresponding element is set to " 1 ", with matrix
Element on the middle leading diagonal is at submatrix H
Lower_rightMiddle corresponding element is set to " 0 ", with mask code matrix
In all values be that the element of " 0 " is at submatrix H
Lower_leftMiddle corresponding element is set to " 1 "; According to the long maximization of minimum ring criterion, be matrix
Element on the middle counter-diagonal is in matrix H
Lower_rightMiddle corresponding element and matrix
The element that middle value is " 1 " is at submatrix H
Lower_leftMiddle corresponding element is provided with corresponding side-play amount;
Six sub-matrices of confirming are merged obtain said basic extension matrix;
Utilize current spreading factor z that said basic extension matrix is carried out correcting process, and utilize big or small replacement matrix P that revised basic extension matrix is expanded, obtain check matrix for z * z, wherein,
Expression rounds up, and K is the information bit length of present encoding, k
b=n
b-m
b
The check matrix that utilization obtains carries out the LDPC coding to data to be encoded.
3. method according to claim 1 is characterized in that,
According to
With
Configuration integer quotient C
1..., C
9, D
1..., D
9, the minimum ring of said basic extension matrix is grown up in equaling 6, wherein, C
1..., C
9Be nonnegative integer, z
MaxBe the maximum of spreading factor,
Be submatrix H
Lower_rightIn the element value of the i ' row j ' row, and i ' and j ' be the index of element on the said counter-diagonal,
Be submatrix H
Lower_leftIn the element value of the capable j of i row, and i and j are submatrix H
Lower_leftIn with mask code matrix
The corresponding element index of element that middle value is " 1 ".
5. according to arbitrary described method in the claim 1 to 4, it is characterized in that, in advance according to a plurality of optional target bit rate select target matrixes and the basic extension matrix under definite each target bit rate;
When carrying out the LDPC coding, according to the corresponding basic extension matrix of current target code rate selection, and carry out said correcting process and extended operation, obtain check matrix.
6. the encoder of the low density parity check code of a rate-compatible is characterized in that, this encoder comprises:
Basic extension matrix is confirmed the unit, is used to select female sign indicating number matrix, and the size (m of definite basic extension matrix
b+ Δ m) * (n
b+ Δ m), wherein, Δ m=Δ xm
b,
Expression rounds up, R
0Be female sign indicating number matrix corresponding code rate, m
bAnd n
bBe respectively the line number and the columns of female sign indicating number matrix, R is a target bit rate; With basic extension matrix H
ExtensionBe divided into six sub-matrices,
Wherein,
Be the information submatrix of said female sign indicating number matrix,
Be the syndrome matrix of said female sign indicating number matrix, H
Upper_right, H
Lower_left, H
Lower_mid, H
Lower_rightSize is respectively m
b* Δ m, Δ m * (n
b-m
b), Δ m * m
b, Δ m * Δ m; With submatrix H
Upper_rightAll elements be set to " 1 "; For submatrix H
Lower_mid, all elements on the preset diagonal is set to " 0 ", and all the other elements are set to " 1 ", and said preset diagonal is for connecting the
Row the 0th column element and nm
b-1 row
The diagonal of column element, and connect (n-1) m
bRow
The element of row and the
Row m
bThe diagonal of-1 column element, wherein, n=1 ..., Δ x, and n is a positive integer; The constructor matrix H
Lower_rightMask code matrix
For possessing many cornerwise lower triangular matrixs, comprise mask code matrix in said many diagonal
Leading diagonal and Δ x bar counter-diagonal, wherein, said Δ x bar counter-diagonal is said matrix H
Lower_midDescribed in preset cornerwise extended line, all elements on said many diagonal is set to " 1 ", all the other elements are set to " 0 "; According to submatrix
H
Upper_right, H
Lower_midWith submatrix H
Lower_rightMask code matrix confirm submatrix H
Lower_leftMask code matrix
Degree distribute the structure mask code matrix
With mask code matrix
In all values be that the element of " 0 " is at submatrix H
Lower_rightMiddle corresponding element is set to " 1 ", with matrix
Element on the middle leading diagonal is at submatrix H
Lower_rightMiddle corresponding element is set to " 0 ", with mask code matrix
In all values be that the element of " 0 " is at submatrix H
Lower_leftMiddle corresponding element is set to " 1 "; According to the long maximization of minimum ring criterion, be matrix
Element on the middle counter-diagonal is in matrix H
Lower_rightMiddle corresponding element and matrix
The element that middle value is " 1 " is at submatrix H
Lower_leftMiddle corresponding element is provided with corresponding side-play amount; Six sub-matrices of confirming are merged obtain said basic extension matrix;
The correcting process unit is used to utilize current spreading factor z that said basic extension matrix is confirmed that the basic extension matrix that the unit obtains carries out correcting process;
Expanding element is used for utilizing big or small replacement matrix P for z * z that the revised basic extension matrix that said correcting process unit obtains is expanded, and obtains check matrix, wherein,
Expression rounds up, and K is the information bit length of present encoding, k
b=n
b-m
b
Coding unit, the check matrix that is used to utilize said expanding element to obtain carries out the LDPC coding to data to be encoded.
7. encoder according to claim 6 is characterized in that, said basic extension matrix confirms that the unit comprises definite subelement and storing sub-units,
Said definite subelement; Be further used for based on a plurality of optional target bit rate select target matrixes and the basic extension matrix under definite each target bit rate, and the target bit rate that each is optional is kept in the said storing sub-units with basic extension matrix is corresponding accordingly;
Said correcting process unit is further used in said storing sub-units, extracting the basic extension matrix corresponding with the current target code check.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200910085516 CN101567697B (en) | 2009-05-25 | 2009-05-25 | Coder and method for coding rate-compatible low-density parity-check codes |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200910085516 CN101567697B (en) | 2009-05-25 | 2009-05-25 | Coder and method for coding rate-compatible low-density parity-check codes |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101567697A CN101567697A (en) | 2009-10-28 |
CN101567697B true CN101567697B (en) | 2012-12-12 |
Family
ID=41283686
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200910085516 Expired - Fee Related CN101567697B (en) | 2009-05-25 | 2009-05-25 | Coder and method for coding rate-compatible low-density parity-check codes |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101567697B (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102412842B (en) * | 2010-09-25 | 2016-06-15 | 中兴通讯股份有限公司 | The coded method of a kind of low density parity check code and device |
WO2013032156A1 (en) | 2011-08-30 | 2013-03-07 | Samsung Electronics Co., Ltd. | Method and apparatus for transmitting and receiving information in a broadcasting/communication system |
KR101922555B1 (en) * | 2011-08-30 | 2018-11-28 | 삼성전자주식회사 | Method and apparatus transmitting and receiving information in broadcasting/communication system |
KR101907943B1 (en) | 2012-01-31 | 2018-12-19 | 삼성전자주식회사 | Apparatus and method for transmitting/receiving data in a communication system |
CN103310251B (en) * | 2012-09-10 | 2015-02-18 | 成都菲普迪斯科技有限公司 | CCQR (Colorful Coded Quick Response) code encoding method and decoding method |
CN105429645B (en) * | 2014-09-17 | 2019-03-08 | 上海数字电视国家工程研究中心有限公司 | For the check matrix, LDPC code word and coding method of low code rate LDPC code |
CN105811996B (en) | 2014-12-30 | 2019-12-06 | 华为技术有限公司 | data processing method and system based on quasi-cyclic LDPC |
WO2018126914A1 (en) * | 2017-01-09 | 2018-07-12 | 中兴通讯股份有限公司 | Method and device for coding of quasi-cyclic low-density parity-check code, and storage medium |
CN108809509B (en) | 2017-05-05 | 2021-01-22 | 电信科学技术研究院 | Method and device for selecting basic diagram of low-density parity check code |
CN109995379B (en) * | 2017-12-29 | 2021-07-16 | 华为技术有限公司 | Method and device for processing information |
CN111464191A (en) * | 2020-05-25 | 2020-07-28 | 重庆邮电大学 | RC-L DPC code construction method based on matrix expansion and Fibonacci sequence |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070038916A1 (en) * | 2005-06-08 | 2007-02-15 | Conexant Systems, Inc. | Method of coding data |
CN1953335A (en) * | 2005-10-21 | 2007-04-25 | 中兴通讯股份有限公司 | A coding device and method for low density parity check code of supporting any code rate/code length |
-
2009
- 2009-05-25 CN CN 200910085516 patent/CN101567697B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070038916A1 (en) * | 2005-06-08 | 2007-02-15 | Conexant Systems, Inc. | Method of coding data |
CN1953335A (en) * | 2005-10-21 | 2007-04-25 | 中兴通讯股份有限公司 | A coding device and method for low density parity check code of supporting any code rate/code length |
Non-Patent Citations (1)
Title |
---|
倪俊枫,甘小莺,张海滨,徐友云.改进的分层修正最小和LDPC译码算法及译码器设计.《系统工程与电子技术》.2008, * |
Also Published As
Publication number | Publication date |
---|---|
CN101567697A (en) | 2009-10-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101567697B (en) | Coder and method for coding rate-compatible low-density parity-check codes | |
CN103152056B (en) | A kind of quasi-cyclic LDPC code constructing method and device based on protograph | |
CN101141133B (en) | Method of encoding structured low density check code | |
EP3457575B1 (en) | Encoding method and device and decoding method and device for structured ldpc | |
CN101162907B (en) | Method and device for constructing low-density parity code check matrix | |
CN101453297B (en) | Encoding method and apparatus for low density generation matrix code, and decoding method and apparatus | |
CN104868925A (en) | Encoding method, decoding method, encoding device and decoding device of structured LDPC codes | |
CN101796488A (en) | Generation of parity-check matrices | |
CN101431337A (en) | Method for improving code parallelism degree and implementing coding delay | |
CN103731160B (en) | Packet space coupling low density parity check coding method | |
CN104917536A (en) | Method and apparatus for supporting low-code-rate encoding | |
CN103053116B (en) | The coded method of low density parity check code and device | |
CN102394659B (en) | Low density parity check (LDPC) code check matrix construction method and corresponding matrix multiply operation device | |
CN101159435B (en) | Low density check code check matrix constructing method based on shift matrix classified extension | |
CN101207386B (en) | Constitution method of binary low density parity check code | |
CN101488760B (en) | Encoding method for low code rate LDPC code | |
CN104967455A (en) | Recursive encoding method of spatially-coupled low-density parity check codes | |
CN102420616B (en) | Error correction method by using quasi-cyclic LDPC code based on Latin square | |
CN101465655B (en) | Method for encoding ultra short code length density parity check code | |
CN101577554B (en) | Method for coding low-density parity check code of multi-code length and multi-code rate | |
CN103199877A (en) | Method for constructing and coding structured LDPC (Low Density Parity Check) convolutional codes | |
CN106685432A (en) | Large girth type-II QC-LDPC code construction method based on complete cyclic difference sets | |
CN102611465B (en) | Coder of structured q-ary irregular repeat-accumulate code and coding method of coder | |
CN105871385A (en) | LDPC convolutional code construction method | |
CN103338044B (en) | Protograph code for deep space optical communication system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121212 |
|
CF01 | Termination of patent right due to non-payment of annual fee |