Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD AND APPARATUS FOR ENCODING AND DECODING OF VARIABLE LENGTH QUASI-CYCLIC LOW-DENSITY PARITY-CHECK, QC-LDPC, CODES
Document Type and Number:
WIPO Patent Application WO/2018/088923
Kind Code:
A1
Abstract:
A method for quasi-cyclic low-density parity-check (QC-LDPC) encoding and decoding of a data packet by a lifted matrix is provided, the method comprising: lifting the QC-LDPC code for maximal code length Nmax and maximal circulant size Z upper of the base matrix; generating a plurality of optimal values r, for a plurality of circulants Z 1, Z 2,....,Z upper based on the QC-LDPC code lifted for maximal length Nmax, 0 ≤ r; ≤ Z upper - 1; saving the generated plurality of optimal values r; corresponding to the plurality of circulants Z,, Z 2,..., Z upper and a matrix for the QC-LDPC code lifted for maximal length N max in the memory unit; receiving a current circulant Z current from the plurality of circulants Z 1, Z 2,....,Z upper; selecting a current optimal value r current from the plurality of optimal values r; stored in the memory unit corresponding to the current circulant Z current; and lifting the base matrix based on the current optimal value r current, wherein a floor scale modular lifting of the base matrix is calculated as: where E(H upper ) is a value of circulant shift in the base matrix for maximal circulant size; wherein 0 ≤ r current ≤ z upper 1 and r current =1 is excluded. The apparatus for QC-LDPC encoding and decoding of a data packet by a lifted matrix is further provided.

Inventors:
USATYUK VASILY STANISLAVOVICH (CN)
VOROBYEV ILYA VIKTOROVICH (CN)
POLIANSKII NIKITA ANDREEVICH (CN)
SVISTUNOV GERMAN VIKTOROVICH (CN)
Application Number:
PCT/RU2016/000777
Publication Date:
May 17, 2018
Filing Date:
November 14, 2016
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
HUAWEI TECH CO LTD (CN)
USATYUK VASILY STANISLAVOVICH (CN)
International Classes:
H03M13/11; H03M13/03
Foreign References:
US20100257425A12010-10-07
Other References:
SEHO MYUNG ET AL: "Lifting Methods for Quasi-Cyclic LDPC Codes", IEEE COMMUNICATIONS LETTERS, IEEE SERVICE CENTER, PISCATAWAY, NJ, US, vol. 10, no. 6, 1 June 2006 (2006-06-01), pages 489 - 491, XP001546983, ISSN: 1089-7798, DOI: 10.1109/LCOMM.2006.1638625
"LDPC coding for OFDMA PHY ; C80216e-05_066r3", IEEE DRAFT; C80216E-05_066R3, IEEE-SA, PISCATAWAY, NJ USA, vol. 802.16e, 27 January 2005 (2005-01-27), pages 1 - 16, XP017625567
MING JIANG ET AL: "An improved variable length coding scheme using structured LDPC codes", PROC., IEEE INTERNATIONAL CONFERENCE ON WIRELESS COMMUNICATIONS AND SIGNAL PROCESSING, WCSP 2010, 21 October 2010 (2010-10-21), pages 1 - 5, XP031800170, ISBN: 978-1-4244-7556-8
Attorney, Agent or Firm:
LAW FIRM "GORODISSKY & PARTNERS" LTD. (RU)
Download PDF:
Claims:
CLAIMS

1. A method (100) for quasi-cyclic low-density parity-check, QC-LDPC, encoding and decoding of a data packet by a lifted matrix, obtained by floor scale modular lifting of a base matrix of a QC-LDPC code,

the method (100) comprising:

lifting (101 ) the QC-LDPC code for a maximal code length NmM and a maximal cir- culant size Zupper of the base matrix, Nmax = Zupper * L , wherein L is the number of columns of the base matrix;

generating (102) a plurality of optimal values r: for a plurality of circulants Zx , Z2 ,..., Zupper based on the QC-LDPC code lifted for maximal length Nmax , 0 < ri ≤ z ^upper - 1

saving (103) the generated plurality of optimal values η corresponding to the plurality of circulants Z, , Z2 Zupper and a matrix for the QC-LDPC code lifted for maximal length Nmax in a memory unit; and

receiving (104) a current circulant Zcurre from the plurality of circulants

Z] , Z2 ,..., Zupper , selecting (105) a current optimal value from the plurality of optimal values η stored in the memory unit corresponding to the current circulant Zcurrenl ; and

lifting (106) the base matrix based on the current optimal value Γ^^^, wherein a floor lifting of the base matrix is calculated as:

Z

\ current / ' z ((EiH upper ) * r current Z upper )

upper

where E(Hupper) is a value of circulant shift in the base matrix-for maximal circulant size;

wherein 0 < rcurrent < Zupper - 1 and rcurrent = 1 is excluded.

2. The method of claim 1 , wherein generating the plurality of optimal values ri for the plurality of circulants Z , Z2 ,..., Z further comprises: constructing a plurality of families of parity-check matrixes, each family corresponds to value r in a plurality of values rv r2,...,rk corresponding to code lengths Nl , N2, N3,..., Nk ; and

based on the plurality of the families of the parity-check matrixes, selecting the plu- rality of optimal values η for the plurality of circulants Z, , Z2,..., Z by multi-parameter filtering.

3. The method of claim 2, wherein the multi-parameter filtering includes at least one of: Extrinsic Message Degree, ACE Spectrum, Tanner Spectral Bound, Code Distance, Codeword's weight spectrum enumerator, Trapping Set Weight Enumerator, simulations result.

4. The method of any of claims 2-3, wherein constructing the plurality of the families of the parity-check matrixes is performed using equation:

^r^u per ~ ^ ( ^ upper) ' T mod i " pper-

5. A computer readable storage medium storing program code, the program code comprising instructions, which when performed on a computer cause the computer to perform the method according to any of claims 1-4. 6. An apparatus (200) for quasi-cyclic low-density parity-check, QC-LDPC, encoding and decoding of a data packet by a lifted matrix, obtained by floor scale modular lifting of a base matrix of QC-LDPC code,

the apparatus (200) comprising a processing unit (201) and a memory unit (202),

the memory unit (202) storing:

a maximal length Nmax and a maximal circulant size Z of the base matrix, a matrix for the QC-LDPC code lifted for maximal length Nmax ; and

a plurality of optimal values rt corresponding to a plurality of circulants

Z, , Z, ,..., Z , the plurality of optimal values η is generated based on the QC-LDPC code lifted for maximal length Nmax and maximal circulant size Z of the base matrix, wherein Nmax = Zupper * L , L is a column in the base matrix and 0 < η ≤ Zupper - 1;

the processing unit (201) configured to:

receive a current circulant Zcurrenl from the plurality of circulants Z, , Z2,..., Z ; select a current optimal value rcurrent from the plurality of optimal values rt stored in the memory unit corresponding to the current circulant Zcurrenl ; and

lift the base matrix based on the current optimal value r^^g^, wherein a floor lifting of the base matrix is calculated as: current ) ' f ccuurrrreennt, / )m°d " Z uuppppeerr ) )

where E{Hupper) is a value of circulant shift in the base matrix for maximal circulant size; wherein 0 < rcurrent < Zupper - 1 and rcurrent = 1 is excluded. 7. The apparatus of claim 6, wherein, generating the plurality of optimal values r( for the plurality of circulants Zx , Z2,..., Zupper comprises:

constructing a plurality of families of parity-check matrixes, each family corresponds to value r in a plurality of values rx , r2 ,...,rk corresponding to code lengths Nl , N2 , Ni,..., Nk ; and

based on the plurality of the families of the parity-check matrixes, selecting the plurality of optimal values r: for the plurality of circulants Z] , Z2,..., Z by multi-parameter filtering.

8. The apparatus of claim 7, wherein the multi-parameter filtering includes at least one of: Extrinsic Message Degree, ACE Spectrum, Tanner Spectral Bound, Code Distance, Codeword's weight spectrum enumerator, Trapping Set Weight Enumerator, simulations result.

9. The apparatus of any of claims 7-8, wherein the processing unit is further configured to construct the plurality of the families of the parity-check matrixes using equation:

Er mod ZUpper.

Description:
METHOD AND APPARATUS FOR ENCODING AND DECODING OF VARIABLE LENGTH QUASI-CYCLIC LOW-DENSITY PARITY-CHECK, QC-LDPC, CODES

TECHNICAL FIELD

The present invention relates to a method for quasi-cyclic low-density parity-check (QC- LDPC) encoding and decoding and an apparatus for quasi-cyclic low-density parity-check encoding and decoding.

The present invention also relates to a computer-readable storage medium storing program code, the program code comprising instructions for carrying out such a method.

BACKGROUND

Error-correcting coding is an efficient method to improve capacity of a communication system. Wireless systems require employing a large set of code with different length and rate. For example LTE provides more than several thousand of different code lengths and rates using a hardware friendly interleaver and simply puncturing pattern, but the sequential nature of the BCJR decoder of Turbo code significantly limits parallelism - decoder throughput.

Hence, it is thus a problem how to create a compact representation of QC-LDPC codes, which supports sets of QC-LDPC codes with different lengths and rates. Other problems to be solved include getting additive increase of circulant size to minimize gap between several lengths of code; to define some block-structured memory efficient puncture pattern with min- imal performance lost; and to maximize number of variable node in block-structured which recover under practical number iteration.

A problem of existing floor lifting methods is the possibility of appearing of short cycles in parity check matrices and bad weight spectrum of codewords. This leads to lower code gain.

SUMMARY OF THE INVENTION

The objective of the present invention is to provide a method for quasi-cyclic low-density parity-check encoding and decoding and an apparatus for quasi-cyclic low-density parity- check encoding and decoding, wherein the method for QC-LDPC encoding and decoding and the apparatus for QC-LDPC encoding and decoding overcome one or more of the above- mentioned problems of the prior art. Aspects of the invention provide error correction, especially to channel coding for wireless communication, such as WI-FI or 5G communication.

The foregoing and other objects are achieved by the features of the independent claims. Further implementation forms are apparent from the dependent claims, the description and the figures. A first aspect of the invention provides a method for quasi-cyclic low-density parity-check (QC-LDPC) encoding and decoding of a data packet by a lifted matrix, obtained by floor scale modular lifting of a base matrix of QC-LDPC code, the method comprising: lifting the QC-LDPC code for maximal code length N max and maximal circulant size Z of the base matrix, N max = * L , wherein L is a column in the base matrix; generating a plurality of optimal values η for a plurality of circulants Z, , Z 2 ,..., Z based on the QC-LDPC code lifted for maximal length N max , 0 < r i < Z upper - 1; saving the generated plurality of optimal values r i corresponding to the plurality of circulants Z ] , Z 2 ,..., Z upper and a matrix for the QC-LDPC code lifted for maximal length N max in the memory unit. These steps may be made offline only once. The method further comprises receiving a current circulant Z currenl from the plurality of circulants Z, , Z 2 Z ; selecting a current optimal value r C urrent fr° m tne plurality of optimal values r i stored in the memory unit corresponding to the current circulant Z arreni '■> an ^ lifting the t> ase matrix based on the current optimal value r current , wherein a floor lifting of the base matrix is calculated as: E(H C y i E ( H u PP er ) * Current )™Od

upper

where E[H pper ) is a value of circulant shift in the base matrix for maximal circulant size; wherein 0 < r current ≤ Z upper - 1 and r^^nt = 1 is excluded. Therefore, a QC-LDPC mother code lifting method with flexible length and rate is provided to be used for encoding and decoding of data packets. This method provides memory efficient QC-LDPC code representation with maximal flexibility of length and rate. The overall code performance is also therefore increased due to providing memory consumption and processing speed increase. The methods according to the first aspect of the invention can be performed by a computer- readable storage medium according to the second aspect of the invention. Further features or implementations of the method according to the first aspect of the invention can perform the functionality of an apparatus for QC-LDPC encoding and decoding according to the third as- pect of the invention and its different implementation forms.

In a first implementation of the method for QC-LDPC encoding and decoding of a data packet by a lifted matrix according to the first aspect, generating the plurality of optimal values η for the plurality of circulants Z, , Z 2 ,..., Z further comprises: constructing a plurality of families of parity-check matrixes, each family corresponds to value r in a plurality of values r r 2 ,...,r k corresponding to code lengths N l , N 2 , N 3 ,..., N k ; and based on the plurality of the families of the parity-check matrixes, selecting the plurality of optimal values r i for the plurality of circulants Z, , Z, Z r by multi-parameter filtering. Using at least one parity- check matrix a set of code can be represented with minimal performance degradation and high memory efficiency.

In a second implementation of the method for QC-LDPC encoding and decoding of a data packet by a lifted matrix according to the first implementation of the first aspect, the multiparameter filtering includes at least on of: Extrinsic Message Degree, ACE Spectrum, Tanner Spectral Bound, Code Distance, Codeword's weight spectrum enumerator, Trapping Set

Weight Enumerator, simulations result. All these ways of choosing best r value provide improved filtering quality due to better consideration of multiple parameters and enable choosing the optimal r value to be used in a lifting procedure. In a third implementation of the method for QC-LDPC encoding and decoding of a data packet by a lifted matrix according to any of the first or second implementations of the first aspect, constructing the plurality of the families of the parity-check matrixes is performed using formula: E r (H upper = E[H upper ) r mod Z upper . Using E r (H upper ) provides additional flexibility due to possibility to choose r value to avoid critical points.

A second aspect of the invention refers to a a computer-readable storage medium storing program code, the program code comprising instructions for carrying out the method of the first aspect or one of the implementations of the first aspect. A third aspect of the invention refers to an apparatus for quasi-cyclic low-density parity-check (QC-LDPC) encoding and decoding of a data packet by a lifted matrix, obtained by floor scale modular lifting of a base matrix of QC-LDPC code, the apparatus comprising a pro- cessing unit and a memory unit, the memory unit storing: a maximal length N max and a maximal circulant size Z upper of the base matrix, a matrix for the QC-LDPC code lifted for maximal length N max ; and a plurality of optimal values r t corresponding to a plurality of circu- lants Z, , Z 2 ,..., Z , the plurality of optimal values r t is generated based on the QC-LDPC code lifted for maximal length N max and maximal circulant size Z of the base matrix, wherein N max = Z upper * L , L is a column in the base matrix and 0 < r i < Z upper - 1. The processing unit is configured to: receive a current circulant Z currenl from the plurality of circulars Z x , Z 2 ,..., Z pper ; select a current optimal value r current from the plurality of optimal values r i stored in the memory unit corresponding to the current circulant Z currem ; and lift the base matrix based on the current optimal value r current , wherein a floor lifting of the base matrix is calculated as : E(H currenl ) = where E{H upper ) is a value of circulant shift in the base matrix for maximal circulant size; wherein 0 <

^current— U pp er 1 and ^current - 1 is excluded.

In a first implementation of the apparatus for QC-LDPC encoding and decoding of a data packet by a lifted matrix of the third aspect, generating the plurality of optimal values r. for the plurality of circulants Z, , Z 2 ,..., Z comprises: constructing a plurality of families of parity-check matrixes, each family corresponds to value r in a plurality of values r x , r 2 ,...,r k corresponding to code lengths N N 2 , N 3 ,..., N k ; and based on the plurality of the families of the parity-check matrixes, selecting the plurality of optimal values r t for the plurality of cir- culants Z, , Z, ,..., Z by multi-parameter filtering.

In a second implementation of the apparatus for QC-LDPC encoding and decoding of a data packet by a lifted matrix according to the first implementation of the third aspect, the multiparameter filtering includes at least on of: Extrinsic Message Degree, ACE Spectrum, Tanner Spectral Bound, Code Distance, Codeword's weight spectrum enumerator, Trapping Set Weight Enumerator, simulations result.

In a third implementation of the apparatus for QC-LDPC encoding and decoding of a data packet by a lifted matrix according to any of the first or second implementations of the third aspect, the processing unit is further configured to construct the plurality of the families of the parity-check matrixes using formula: E r (H upper ) = E(H upper ) r mod Z upper .

All the implementations of the first aspect may be easily combined and used together with all the implementations of the third aspect.

These and other aspects of the invention will be apparent from the embodiments described below.

BRIEF DESCRIPTION OF THE DRAWINGS

To illustrate the technical features of embodiments of the present invention more clearly, the accompanying drawings provided for describing the embodiments are introduced briefly in the following. The accompanying drawings in the following description are merely some embodiments of the present invention, modifications on these embodiments are possible without departing from the scope of the present invention as defined in the claims.

FIG. 1 is a flow chart of a method for QC-LDPC encoding and decoding of a data packet by a lifted matrix in accordance with an embodiment of the present invention,

FIG. 2 shows generating a lifting value r for floor-scale modular lifting method in accordance with the present invention, FIG. 3 shows an example of critical points elimination due to changing r values, FIG. 4 is a block diagram illustrating an apparatus for QC-LDPC encoding and decoding of a data packet by a lifted matrix in accordance with an embodiment of the present invention, FIGS. 5-7 show comparison of the floor lifting in accordance with the present invention with traditional floor lifting.

Detailed Description of the Embodiments FIG. 1 illustrates a method 100 for QC-LDPC encoding and decoding of a data packet by a lifted matrix in accordance with the first aspect of the invention. The lifted matrix is obtained by floor scale modular lifting of a base matrix of QC-LDPC code. The method starts at block 101 , where the QC-LDPC code for maximal code length N max and maximal circulant size

^ upper °f me b ase matrix is lifted.

N ms = Z upper * L , (1) wherein L is a column in the base matrix.

At step 102 a plurality of optimal values η for a plurality of circulants Z, , Z, Z upper is generated based on the QC-LDPC code lifted for maximal length N max , 0 < r t ≤ Z upper - 1. The generated plurality of optimal values r ; corresponding to the plurality of circulants

Z, , Z 2 Z and a matrix for the QC-LDPC code lifted for maximal length N max are saved in the memory unit at step 103. At step 104 a current circulant Z curre from the plurality of circulants Z, , Z 2 ,..., Z is received. Then a current optimal value r current is selected from the plurality of optimal values η stored in the memory unit corresponding to the current circu- lant Z currem (step 105). Finally at step 106 the base matrix is lifted based on the current optimal value r current . A floor lifting of the base matrix is calculated as:

7

current \

E^JEl current } ' - ((E{H upper )* r currenl )mod Z upper ) | , (2)

7 upper

where E[H ) is a value of circulant shift in the base matrix for maximal circulant size; wherein 0≤ r current ≤ Z upper - 1 and r current = 1 is excluded. The method for QC-LDPC encoding and decoding in accordance with the present invention may be widely used, for example in cryptography, in data transfer and for data storage.

A (J,L) regular QC-LDPC code of length N is usually defined by a parity-check matrix

where 1 < / < J - 1 , 1 < / < Z - 1 and l{p , ) represents the p x p circulant permutation matrix obtained by cyclically right-shifting the px p identity matrix l(o) by p , positions, with p = N/ L. For a specific QC-LDPC code the corresponding "base matrix" ("mother matrix" or protograph) is defined as the matrix of circulant shift that defines the QC-LDPC code:

Mask matrix for which regular QC-LDPC code can become irregular for different column weight case or QC-LDPC regular code with zero block circulant may be defined as:

H = H® , (6) where ® is Hadamard product.

Lifting is operation under base matrix (protograph), by using of which code with different authomorphism or circulant size from a similar base matrix can be obtained.

Normally floor-lifting of base matrix is calculated by formula:

where z currenl - lifting size of circulant,

z upper - maximal circulant size of base matrix,

E [H upper) - value of circulant shift in base matrix for maximal size of circulant.

Length of code N from z currenl * VN prolograph to z upper * VN protograph with some additive step between z current ' - ste P '∑ upper , where VN prolograph is number of variable nodes in base matrix.

The method according to the present invention uses random matrix design approach lifting QC-LDPC directly from mask matrix (base matrix or protograph).

A cycle of even length 2K in H is defined by 2K positions such that:

1 ) Two consecutive positions are obtained by changing alternatively column of row only;

2) All positions are distinct except first and last one;

Two consecutive elements of the path belong to different circulant permutation matrices. So a chain of circulant permutation matrices can be defined:

/ iPi 0 , jo ). l iPiojJ. I (PH ' I iP . ) iPi K → K-J> 1 (PiK-iJo)' 7 (Pio Jo) ( 8 ) where i a ≠ i a + \ ,i a ≠ ja+ τ all 0 < a≤ K - 1.

As each part of cycle is one, the circulant permutation matrix /(p ) participating in cycle cannot be empty. Using these shifts of identity matrix necessary and sufficient conditions of existing of the cycle can be defined as:

Fig. 2 illustrates generating a lifting value r for floor-scale modular lifting method. To gener- ate a lifting value (code for flexible length N, < N 2 < N 3 < N < ... < N k < N max ) using floor scale modular approach QC-LDPC lifted for maximal length N max is used. This matrix can be lifted using simulation annulling, hill-climbing, guest-and-search, PEG, ACE+PEG or any another algorithms. 0,0 m, o.i m Ο. , -1

m 1,0 m. m. Ι , -1

M = (10)

The input for floor modular scale lifting represents a QC-LDPC code lifted for maximal circu- lant size Z upper with L variable nodes (columns in base matrix) and J parity-check (rows in base matrix):

^ upper * L = N max . (1 1)

Circulant sizes for which lifting this base matrix is desired: Ζ < Z 2 < ... < Z k < Z upper , to get lengths Z, * L = N, < Z 2 * L = N 2 < Z, * L = N 3 < Z 4 * L = N 4 < ... < Z upper * L = N max . The output of the floor modular scale lifting represents scale values r^ r 2 ,...,r k for every circulant sizes Z, , Z 2 ,...,Z k . By using these values it is possible to generate code for every code lengths N , N 2 , N i ,...,N k in a fast manner using formula (12).

QC-LDPC code lifted for maximal code length is received and for every value r^ r 2 ,...,r k (related to circulants size Z x , Z 2 ,...,Z k ) which corresponds to codes with lengths

N N 2 , N } ,...,N k using formula (12) i parity-check matrixes are determined. Every r cwrenl can be in the range 1...Z, - 1 . After using multi-parameter sieving, best value r is chosen based on: Extrinsic Message Degree, ACE Spectrum, Tanner Spectral Bound, Code Distance, Codeword's weight spectrum enumerator, Trapping Set Weight Spectrum Enumerator, simulation result, as shown in Fig. 2 by ACE=EMD Analysis. This procedure may be made offline only once, then a matrix lifted for maximal length is saved along with r values.

After value r ( is got for every circulant Z, , Z 2 Z upper parity-check matrix for every length N N 2 , N3 ,...,N k can be constructed using formula: ^r^^upper) = ^ ( ^ ^ upper) ' T mod ^u per 0¾ where r is integer 1≤ r≤ z U pp er - 1 and GCD{r, z upper ) = 1. For any path P shift d P ' in the E r [H upper ) is equal to r times of shift d P by same path in E[H upper )

K-l

d'p≡ rp ia a - rp iaJa+i (mod z upper ) ≡

a-0

(13)

K-l

PiaJa - Pia a+i ( m ° d Z upper)≡ T d f

a=0

When GCD(r, z upper ) = 1, d P ' = 0 {mod z upj)er ) in the same time with d P

0 (mod z upper ).

So, structure of cycles of E[H upper ) and E_r(H upper ) are equivalent. In comparison with classical floor lifting approach, this method provides additional freedom and flexibility. Such r can be chosen as to avoid catastrophic (critical) points and to improve quality of graph in general, example of improving using change of r is presented in Fig 3.

Combining formula (12) with formula (7) of classical floor-lifting of base matrix the following formula for a floor lifting can be obtained: where r current - scale factor, being an integer value from 0... z upper - \ C GD rcurrent' z upper) ~ 1 ·

This increases freedom and flexibility of floor lifting.

For each z current we can find such r current that bring best possible quality of E(H current ). This method can be applied to any QC-LDPC codes to get flexible length properties.

Fig. 4 illustrates an apparatus 200 for QC-LDPC encoding and decoding of a data packet by a lifted matrix comprising a processing unit 201 and a memory unit 202. The memory unit 202 stores: a maximal length N MAX and a maximal circulant size Z of the base matrix, a matrix for the QC-LDPC code lifted for maximal length N MAX ; and a plurality of optimal values r t corresponding to a plurality of circulants Z, , Z 2 Z . The plurality of optimal values r i is generated based on the QC-LDPC code lifted for maximal length N max and maximal circu- lant size Z upper of the base matrix. The processing unit 201 is configured to: receive a current circulant Z currcnl from the plurality of circulants Z, , Z 2 ,..., Z WCT . ; select a current optimal value rcurrent fr° m the plurality of optimal values η stored in the memory unit corresponding to the current circulant Z currenl ; and lift the base matrix based on the current optimal value r current .

Comparison of the floor lifting in accordance with the present invention with traditional floor lining is further provided with the reference to Figs. 5-7. For simplicity of the comparison, the extended irregular repeat-accumulate (EIRA) QC-LDPC base matrix is lifted. Designed QC- LDPC is:

Table 1 contains a comparison of the lifting approach in accordance with the provided method and a traditional floor lifting approach based on the number of cycles.

Floor scale modular lifting in acTraditional floor lifting

cordance with provided method

•7

" current cycles Number of - current cycles Number of

cycles; r cycles; r

24 6 669 ; r = 4 24 4 2 r = 1

28 6 583 r = 28 28 4 2 r = 1

32 6 508 r = 52 32 4 1 r = l

36 6 439 r = 20 36 4 1 r = 1

40 6 420 r = 28 40 4 1 r = 1

44 6 364 r = 76 44 4 1 r = 1

48 6 238 ; r = 2 48 4 1 r = 1

52 6 341 r = 20 52 4 1 r = 1

56 6 217 r = 14 56 4 1 r = 1

60 6 195 r = 10 60 4 1 r = 1

64 6 179 r = 74 64 6 263 ; r = 1

68 6 177 r = 86 68 6 223 ; r = l

72 6 151 r = 58 72 6 228 ; r = 1

76 6 153 r = 94 76 6 223 ; r = 1

80 6 144 r = 14 80 6 205 ; r = 1 84 6 139 ; r = 82 84 6 201 ; r = 1

88 6 130 ; r = 38 88 6 197; r = 1

92 6 132 ; r = 86 92 6 178 ; r = 1

2 upper = 9£> 6 68 ; r = 1 - pper 96 6 173 ; r = 1

Table 1

Comparison based on the ACE Spectrum for lifting with circulant size z cu enl = 60, N=1440 is provided in Fig.5, where ACE Spectrum of floor scale modular lifting in accordance with the provided method (left) and traditional floor lifting (right) under similar base matrix are provided.

BER performance comparison of traditional floor-lifting QC-LDPC and QC-LDPC lifted using the provided approach of same base-matrix under min-sum decoder 15 iterations under AWGN channel is illustrated in Fig.6. FER performance comparison of traditional floor- lifting QC-LDPC and QC-LDPC lifted using the provided approach of same base-matrix under min-sum decoder 15 iterations under AWGN channel is illustrated in Fig.7.

Using the floor scale modular lifting method described in the present description the follow- ing two parity-check matrix of Repeat Accumulate QC-LDPC code may be designed:

12x24 circulant from 28 to 2304 with step 4, length 672 to 55296 with step 96, rate 0.5

and

6x24 circulant from 4 to 2304 with step 4 length 96 to 55296 with step 96, rate 0.75

984 581 2108 942 855 1987 1404 -1 1365 -1 2025 -1 -1 667 719 804 -1 -1 899 0 -1 -1 -1 -1

682 737 1893 932 2126 185 1472 522 -1 377 1122 1161 -1 -1 -1 -1 934 212 -1 0 0 -1 -1 -1

83 1971 342 858 1726 2205 815 -1 -1 109 671 -1 1876 442 -1 447 1576 -1 0 -1 0 0 -1 -1

201 907 1490 191 272 1986 970 616 1393 -1 -1 646 -1 -1 930 270 -1 629 -1 -1 -1 0 0 -1

2158 2244 1820 390 1445 2051 861 -1 1454 1022 1163 -1 139 742 -1 -1 - 1 -1 -1 -1 -1 -1 0 0

679 421 874 2035 1806 723 2097 884 -1 -1 -1 19 1449 -1 1793 -1 1081 275 899 -1 -1 -1 -1 0 The foregoing descriptions are only implementation manners of the present invention, the scope of the present invention is not limited to this. Any variations or replacements can be easily made through person skilled in the art. Therefore, the protection scope of the present invention should be subject to the protection scope of the attached claims.